Quaternary Incision, Salt Tectonism, and Landscape Evolution of Moab-Spanish Valley, Utah

Home , La Sal Mountains, Moab Fault, Mount Mellenthin

Utah State University DigitalCommons@USU

All Graduate Theses and Dissertations Graduate Studies

5-2018

James P. Mauch Utah State University

Follow this and additional works at: https://digitalcommons.usu.edu/etd

Part of the Geology Commons

Recommended Citation Mauch, James P., "Quaternary Incision, Salt Tectonism, and Landscape Evolution of Moab-Spanish Valley, Utah" (2018). All Graduate Theses and Dissertations. 7068. https://digitalcommons.usu.edu/etd/7068

This Thesis is brought to you for free and open access by the Graduate Studies at DigitalCommons@USU. It has been accepted for inclusion in All Graduate Theses and Dissertations by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. QUATERNARY INCISION, SALT TECTONISM, AND LANDSCAPE

EVOLUTION OF MOAB-SPANISH VALLEY, UTAH

James P. Mauch

A thesis submitted in partial fulfillment of the requirements for the degree

MASTER OF SCIENCE

Geology

Approved:

______Joel L. Pederson, Ph.D. Tammy M. Rittenour, Ph.D. Major Professor Committee Member

______Susanne U. Jänecke, Ph.D. Mark R. McLellan, Ph.D. Committee Member Vice President for Research and Dean of the School of Graduate Studies

UTAH STATE UNIVERSITY Logan, Utah

Quaternary Incision, Salt Tectonism, and Landscape

Evolution of Moab-Spanish Valley, Utah

James P. Mauch, Master of Science

Utah State University, 2018

Major Professor: Dr. Joel L. Pederson Department: Geology

Modern topography and fluvial terraces can be used to understand deformation and incision. These approaches are applied to the Moab-Spanish Valley salt graben in the central Colorado Plateau. Here, localized salt deformation is superimposed on regional incision, yet the co-evolution of these processes is not understood. Using mapping, luminescence and cosmogenic dating, surveying, and geographic information system

(GIS) terrain analyses, this work seeks to understand the rates, patterns, controls, and geomorphic signatures of the fluvial and salt-tectonic processes that have sculpted this landscape.

Deposits of Colorado River tributaries Mill and Pack creeks are offset along fault zones and down-warped along the Moab-Spanish Valley graben axis. Luminescence dating and surveying constrains displacement rates over the last 100,000 years to approximately 440 meters per million years (m/my) at the northeast graben margin across discontinuous, steeply-dipping faults with atypical geometries. Over the same time, total iv subsidence has occurred at about 450 m/my in Spanish Valley and up to 1000-1350 m/my

in Moab Valley. Late Pleistocene incision of Mill Creek has kept pace with northeast

graben-margin fault slip, though chronostratigraphy indicates incision prior to 200,000

years ago was much slower. Meanwhile, digital projection of cosmogenically-dated upland gravels indicates that subsidence likewise must have accelerated in the middle- late Pleistocene and that graben collapse is essentially a Quaternary process. Graben subsidence accelerating in conjunction with fluvial incision supports the conceptual model for deformation, where salt-dissolution and transport by groundwater drives subsidence.

GIS-derived indices of normalized steepness and concavity demonstrate that longitudinal profiles of regional tributaries are primarily controlled by heterogeneous bedrock. Yet common baselevel signals are also evident in the distribution of major knickzones in tributaries both crossing and independent of the graben recorded by the Chi parameter. This suggests drainages are responding to regional transient incision of the

Colorado River system and that local subsidence is largely compensated by depositional filling of the graben. This study adds to the growing number to document late Pleistocene transient incision across the central Colorado Plateau. Such results require a baselevel fall downstream, either much-delayed from the integration of the Colorado River or from a more recent, unknown source.

(253 pages) v PUBLIC ABSTRACT

Quaternary Incision, Salt Tectonism, and Landscape

Evolution of Moab-Spanish Valley, Utah

James P. Mauch

To study the history of processes that shape the Earth’s surface, geologists look for markers in the landscape that they can date and use to measure change. Rivers leave such markers in their deposits and terrace landforms and in the overall shape of their elevation profile from head to toe. This thesis uses luminescence and cosmogenic methods to date the sediment in terraces to determine when the river deposited it. Field mapping and global positioning system (GPS) surveying are also used to measure the distance between terrace levels to quantify how much change has occurred. This study seeks to answer questions about when, how quickly, and why streams near Moab in

Utah’s Canyonlands region have incised into bedrock. It seeks also to determine the history, rates, and patterns of the active and incremental sinking of Moab and Spanish valleys.

The results from this work show that the incision of canyons along Mill and Pack creeks upstream from Moab accelerated around 200,000 years ago to between 0.44 and

0.62 millimeters per year (mm/yr). The sinking of Moab and Spanish valleys also appears to have quickened around the same time and has occurred at up to 1.35 mm/yr in Moab

Valley and around 0.45 mm/yr in Spanish Valley over the last 100,000 years. The

Kayenta Heights fault zone (KHFZ) accommodates part of the sinking of Moab Valley and has slipped at an average rate of 0.44 mm/yr in the last 100,000 years. This study’s vi mapping of the length and width of individual fault strands supports prior interpretations

that the KHFZ does not pose an earthquake hazard, though it can still be expected to

produce active rock fall and rupture the ground surface locally.

That the sinking of Moab and Spanish valleys and the down-cutting of the

upstream canyons has occurred at similar rates, and may indeed have sped up at a similar

time in the past, indicates a linkage of the two processes. This is consistent with an

existing hypothesis that Moab and Spanish valleys are sinking because groundwater is

dissolving the buried salt deposits and transporting them out through the Colorado River.

This process is able to continue because as the Colorado River and its tributaries incise lower into the landscape, groundwater follows and can reach greater depths in the salt deposits. The results of this study paint a picture of the recent and accelerated sinking of

Moab and Spanish valleys, which has occurred jointly with incision of the regional drainage network. This acceleration of incision, which has been documented throughout the Canyonlands region, may relate to the Colorado River’s carving of Grand Canyon around 6 million years ago or may have come from a more recent and unknown downstream source. vii ACKNOWLEDGMENTS

I am grateful for the generosity of the organizations that provided financial support for this work. These include the National Science Foundation Graduate Research

Fellowship Program, the U.S. Geological Survey EDMAP Program, EarthScope, the

Geological Society of America, the Society for Sedimentary Geology, the Utah State

University (USU) Office of Research and Graduate Studies, and the USU Geology

Department.

This work has been a joint effort that received important contributions from many people who have been generous with their time and went above and beyond to help me succeed.

I was lucky to be accompanied and assisted during dawn-to-dusk days in the field by Jon Harvey, Sherm Young, Zac Williams, Rob McDermott, Mike Turley, and Tim

Garvey. Carlie Ideker and Michelle Nelson at the USU Luminescence Lab were timely and helpful in the processing and analysis of luminescence samples. Alan Hidy and Susan

Zimmerman at the Center for Accelerator Mass Spectrometry at Lawrence Livermore

National Lab provided abundant hospitality, mentorship, and teaching during my 10-day visit to process and analyze cosmogenic burial samples. Rob McDermott deserves another acknowledgement for helping overcome hurdles in the longitudinal profile analysis workflow at a critical juncture, as well as for his efforts to bring the USU

Geology Department’s Vr workstation up to speed. Grant Willis and Kent Brown at the

Utah Geological Survey (UGS) provided abundant mentorship and technical support throughout my mapping project (Chapter 4). I would not have been able to produce a viii complete map product without their tireless assistance. Jessica Castleton and Ben

Erickson at the UGS shared fruitful conversations and insight about geologic hazards

around Moab.

Thank you to committee members Dr. Susanne Jänecke and Dr. Tammy Rittenour

for being patent when I missed deadlines, responsive when I needed feedback, and

supportive throughout the entire research and writing process. Their collective input has

made my thesis much stronger than it would otherwise be. I owe a sincere and special

gratitude to my advisor Dr. Joel Pederson, who has been an unwavering advocate since I

first stepped foot in Logan. His patience, guidance, expertise, and friendship have been

exactly what I’ve needed through these years, and I am lucky to have been his student. I

walk away from my time under his mentorship as a better geologist and person, with a

renewed commitment to remain rooted in the landscapes that inspire me.

Thank you as well to all of the friends and skiing/biking partners who have kept

my life balanced while in graduate school – my time in Logan has been greatly enriched because of your companionship. Finally, I owe a special thank-you to my parents, Joe and Liz Mauch. They have been a steadfast and irreplaceable support system throughout my entire educational journey. Looking back as I now near the finish line, it is clear that I owe all success to their love and support.

James P. Mauch

ix CONTENTS

Page

ABSTRACT ...... iii

PUBLIC ABSTRACT ...... v

ACKNOWLEDGMENTS ...... vii

LIST OF TABLES ...... xiii

LIST OF FIGURES ...... xv

CHAPTER

1. INTRODUCTION ...... 1

2. QUATERNARY SALT TECTONICS AND TRANSIENT INCISION OF THE MOAB-SPANISH VALLEY REGION RECORDED BY CHRONOSTRATIGRAPHY OF DEFORMED ALLUVIAL DEPOSITS ..... 4

ABSTRACT ...... 4

INTRODUCTION ...... 5

BACKGROUND ...... 8

Cenozoic uplift and incision of the Colorado Plateau ...... 8 Salt deformation in the central Colorado Plateau ...... 10 Geologic setting and previous Quaternary research in the Moab- Spanish valley area ...... 13 Fluvial terraces as geomorphic markers and their application to incision of the Colorado Plateau ...... 16 Geochronology ...... 18

METHODS ...... 20

Mapping ...... 20 Surveying ...... 20 Luminescence geochronology ...... 21 Cosmogenic radionuclide geochronology ...... 22 Rate calculations ...... 23

RESULTS ...... 25

Overview ...... 25 x Distribution, geometry, and chronostratigraphy of Mill Creek deposits ...... 26

Mill Creek terraces T1-T5 ...... 28 Mill Creek upland gravels T6 and T7 ...... 33 Mill Creek undifferentiated alluvial gravel deposits ...... 35

Distribution, geometry, and chronostratigraphy of Pack Creek deposits ...... 37

Pack Creek terraces T1-T3 ...... 37 Pack Creek upland gravel T7 ...... 42 Pack Creek undifferentiated alluvial gravel deposits ...... 43

Incision and deformation rates ...... 44

DISCUSSION ...... 53

Controls and correlations of alluvial deposits ...... 53 Graben subsidence and faulting informed by deformed fluvial terraces ...... 58 Relief production in the Moab-Spanish Valley salt graben ...... 60 Controls on the subsidence of Moab and Spanish valleys ...... 62 Incision rates and implications for regional landscape evolution ...... 64 Summary ...... 67

REFERENCES ...... 68

3. TOPOGRAPHIC INSIGHTS INTO SALT TECTONICS AND INCISION AROUND MOAB AND SPANISH VALLEY, UTAH ...... 78

ABSTRACT ...... 78

INTRODUCTION ...... 79

BACKGROUND ...... 82

Cenozoic uplift and incision of the Colorado Plateau ...... 82 Salt deformation in the central Colorado Plateau ...... 86 Geomorphic responses to baselevel fall ...... 86 River longitudinal profiles and bedrock controls...... 87 Topographic indices ...... 89

METHODS ...... 93

Longitudinal profiles ...... 93 Normalized steepness and concavity indices ...... 96 xi Chi analysis ...... 99 Reconstruction of relict topography...... 100

RESULTS ...... 102

Longitudinal profiles and knickzones ...... 102 Normalized steepness and concavity ...... 107 Chi analysis ...... 112 Paleotopography reconstruction ...... 115

DISCUSSION ...... 118

Geologic influences on longitudinal profile form ...... 118 Chi analysis: Implications for regional transience and baselevel fall ...... 122 Reconciling transient incision and bedrock control ...... 124 Paleotopographic and geochronologic constraints on graben evolution ...... 127 Summary ...... 132

REFERENCES ...... 133

4. GEOLOGIC MAP OF THE SOUTHERN HALF OF THE RILL CREEK AND NORTHERN HALF OF THE KANE SPRINGS QUADRANGLES, UTAH ...... 141

ABSTRACT ...... 141

INTRODUCTION ...... 142

SETTING ...... 144

Geographic setting ...... 144 Geologic setting ...... 145

QUATERNARY STRATIGRAPHY ...... 149

Alluvial deposits ...... 150 Colluvial deposits...... 156 Eolian deposits ...... 156 Mass-movement deposits ...... 157 Disturbed deposits ...... 159

STRUCTURE/NEOTECTONICS ...... 160

Moab fault zone ...... 160 Kayenta Heights fault zone ...... 165 xii Implications for graben evolution and subsurface geometry ...... 169

GEOLOGIC HAZARDS ...... 173

Faulting and seismicity ...... 173 Rock fall and other mass movement ...... 175 Ground subsidence ...... 177

REFERENCES ...... 178

5. SUMMARY ...... 183

Chronostratigraphy of deformed alluvial deposits ...... 183 Topographic insights into baselevel, substrate control, and active salt tectonism ...... 185 Mapping of Spanish Valley ...... 186 Practical implications for the city of Moab...... 188

REFERENCES ...... 190

APPENDICES ...... 192

Appendix A: Geochronology data ...... 193

Luminescence data ...... 194 OSL probability density functions and radial plots for equivalent dose distribution ...... 198 IRSL probability density functions and radial plots for equivalent dose distribution ...... 202 Cosmogenic radionuclide (CRN) data ...... 203

Appendix B. GPS survey data ...... 205

GPS survey details ...... 206

Appendix C: Supplemental terrain analysis data ...... 230

Overview ...... 231 Python batch-processing script for vertically-sampled longitudinal profiles and slope-area plots...... 233

xiii LIST OF TABLES

Table Page

2.1 FINAL LUMINESCENCE GEOCHRONOLOGY FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS ...... 27

2.2 PRELIMINARY COSMOGENIC NUCLIDE GEOCHRONOLOGY FOR T7 DEPOSITS ...... 28

3.1 KNICKPOINTS SUMMARY ...... 106

3.2 CONCAVITY, STEEPNESS, AND RELIEF STATISTICS ...... 109

3.3 LONGITUDINAL PROFILE PROJECTION DATA ...... 113

3.4 PALEOTOPOGRAPHIC DIFFERENCING STATISTICS ...... 120

A.1 FINAL LUMINESCENCE GEOCHRONOLOGY FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS ...... 194

A.2 SUMMARY OF LUMINESCENCE DOSE RATE DATA FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS ...... 195

A.3 GRAIN SIZE PROPORTIONS AND α, β AND γ DOSE RATE INFORMATION ...... 197

A.4 PRELIMINARY COSMOGENIC RADIONUCLIDE GEOCHRONOLOGY FOR T7 DEPOSITS ...... 203

A.5 CRN SAMPLE SITE INFORMATION...... 203

A.6 CRN AMS DATA ...... 203

B.1 SURVEY DATA FOR MILL1 GPS TRANSECT (RK 5.0) ...... 208

B.2 SURVEY DATA FOR MILL2 GPS TRANSECT (RK 5.7-5.0) ...... 210

B.3 SURVEY DATA FOR MILL3 GPS TRANSECT (RK 6.3) ...... 212

B.4 SURVEY DATA FOR MILL4 GPS TRANSECT (RK 9.3) ...... 215

B.5 SURVEY DATA FOR MILL5 GPS TRANSECT (RK 19.0) ...... 217

B.6 SURVEY DATA FOR PACK1 GPS TRANSECT (RK 20.7) ...... 220

B.7 SURVEY DATA FOR PACK2 GPS TRANSECT (RK 21.9) ...... 222 xiv B.8 SURVEY DATA FOR PACK3 GPS TRANSECT (RK 23.5) ...... 224

B.9 SURVEY DATA FOR PACK4 GPS TRANSECT (RK 25.4) ...... 228

C.1 SUPPLEMENTAL CATCHMENT STATISTICS ...... 231

xv LIST OF FIGURES

Figure Page

2.1 Location of Moab and Spanish valleys and Colorado River tributaries Mill and Pack creeks ...... 7

2.2 Regional map of the ancestral Paradox Basin ...... 11

2.3 Location of topographic survey transects along Mill and Pack creeks ...... 21

2.4 Mapped distribution of alluvial deposits and faults along Mill Creek at the edge of Moab Valley ...... 29

2.5 Terraces and knickpoint at the Mill Creek trailhead ...... 30

2.6 Mill Creek valley cross sections ...... 31

2.7 Johnson’s Ridge CRN sampling location ...... 35

2.8 Map of alluvial deposits in Pack Creek and southeast Spanish Valley ...... 38

2.9 Pack Creek valley cross sections ...... 40

2.10 Oblique photo of Pack Creek fan in southeast Spanish Valley ...... 41

2.11 South Mountain CRN sampling location ...... 43

2.12 Gravel pit exposure of Spanish Valley upper undifferentiated gravels ...... 45

2.13 Mill Creek incision history ...... 46

2.14 Mill Creek longitudinal profile, incision, and deformation diagram ...... 48

2.15 Incision history at the Pack Creek/Brumley Creek confluence (RK 26.8) ...... 51

2.16 Pack Creek longitudinal profile and terrace deformation diagram ...... 54

3.1 Map of the Moab-Spanish Valley watershed and study tributaries ...... 81

3.2 Regional map of the central Colorado Plateau and ancestral Paradox Basin ...... 84

3.3 Longitudinal profiles and slope-area plots for study tributaries ...... 104 xvi 3.4 Normalized channel steepness map (θref = 0.35) and knickpoint locations ...... 111

3.5 Knickpoints plotted by elevation and classification ...... 112

3.6 Cumulative Chi plots relative to a Colorado River datum ...... 114

3.7 Cumulative Chi plots relative to a graben floor datum ...... 116

3.8 Regional Chi maps ...... 117

3.9 Paleotopographic reconstruction of piedmonts above Moab (MV) and Spanish (SV) valleys ...... 119

4.1 Map area for the southern half of the Rill Creek and the northern half of the Kane Springs 7.5’ quadrangles ...... 143

4.2 Regional map of the ancestral Paradox Basin ...... 146

4.3 Oblique photo of Pack Creek terraces in southeast Spanish Valley ...... 153

4.4 Undifferentiated alluvium (Qagu) along Mill Creek at the edge of Moab Valley ...... 155

4.5 Fissure-fill/collapse breccia along the northeast graben margin below Johnson’s Ridge ...... 159

4.6 Master and subsidiary faults in the Moab fault zone ...... 162

4.7 Kayenta Heights fault zone deformation textures ...... 167

A.1 OSL probability density functions and radial plots for equivalent dose distribution ...... 198

A.2 IRSL probabililty density functions and radial plots for equivalent dose distribution ...... 202

A.3 Johnson’s Mesa CRN isochron ...... 204

A.4 South Mountain CRN isochron ...... 204

B.1 GPS transect locations ...... 207

B.2 Mill1 GPS transect (RK 5.0) ...... 208

B.3 Mill2 GPS transect (RK 5.7-5.0) ...... 210

B.4 Mill3 GPS transect (RK 6.3) ...... 212 xvii B.5 Mill4 GPS transect (RK 9.3) ...... 215

B.6 Mill5 GPS transect (RK 19.0) ...... 217

B.7 Pack1 GPS transect (RK 20.7) ...... 220

B.8 Pack2 GPS transect (RK 21.9) ...... 222

B.9 Pack3 GPS transect (RK 23.5) ...... 224

B.10 Pack4 GPS transect (RK 25.4) ...... 227

C.1 Slope distribution histograms ...... 232

Plate 1 Geologic map of the southern half of the Rill Creek and northern half of the Kane Springs 7.5’ quadrangles ...... in pocket

Plate 2 Supporting information; Geologic map of the southern half of the Rill Creek and northern half of the Kane Springs 7.5’ quadrangles ...... in pocket

Plate 3 Supplementary map; Geologic map of the southern half of the Rill Creek and northern half of the Kane Springs 7.5’ quadrangles ...... in pocket

CHAPTER 1

INTRODUCTION

Streams provide markers of deformation through their deposits and longitudinal

profiles. In the ancestral Paradox Basin of the central Colorado Plateau, fluvial terraces

and longitudinal profiles of Colorado River tributaries record localized Quaternary salt

tectonism, which is superimposed on broader-scale regional incision. Though this scenic

erosional landscape is well known for its active salt tectonism and the canyon-cutting of

the Colorado River, the rates, patterns, and coevolution of these processes is not

understood. This research addresses questions about the rates, timing, and controls of

landscape evolution in the Moab-Spanish Valley salt graben and nearby Colorado River

tributaries in eastern Utah. The results highlight a youthful landscape shaped by the

linked processes of late Quaternary salt deformation and transient fluvial incision.

This thesis consists of three related studies set in and around the Moab-Spanish

Valley salt graben. The first study is detailed in Chapter 2 and develops a

chronostratigraphy of deformed alluvial deposits along the Colorado River tributaries of

Mill and Pack creeks through luminescence and cosmogenic burial dating, topographic

surveying with GPS, and field mapping. This chronostratigraphy is used to determine the

rates, timing, and spatial patterns of salt-dissolution subsidence, graben-margin faulting, and fluvial incision. The results document fluvial terraces and valley-floor alluvium deformed by late-Pleistocene graben subsidence and faulting, incision rates that have accelerated in the late-Pleistocene, and early-Pleistocene upland gravels that record a period of baselevel stability preceding graben collapse. This chapter is a draft manuscript intended for submission to the Geological Society of America Bulletin. 2 Chapter 3 is a geographic information system (GIS) study of topography and

longitudinal profiles of six Colorado River tributaries of the region surrounding the

Moab-Spanish Valley salt graben. Using quantitative metrics from longitudinal profiles

and paleotopographic reconstruction, this work analyzes the evolution of regional

topography in response to baselevel signals and bedrock strength. The results

demonstrate the difficulty in distinguishing baselevel signals from the primary bedrock

control of this erosional landscape. Yet longitudinal profiles still contain subtle baselevel signals in the distribution of knickzones and knickpoints that track transient incision from the Colorado River propagating across Moab and Spanish valleys into the upper reaches of tributary drainages. This chapter is a draft manuscript intended for submission to the journal Earth Surface Processes and Landforms.

Chapter 4 is a mapping study of the northern half of the Kane Springs 7.5’ quadrangle and the southern half of the Rill Creek 7.5’ quadrangle, an area east of Moab,

Utah that encompasses most of Spanish Valley. The purpose of this mapping is to

determine the spatial patterns of deformed Quaternary deposits, assess the style, geometry, and extent of graben-bounding structures, and document the bedrock geology outside of the graben. The chapter includes the complete 1:24,000 scale map (Plate 1),

unit descriptions, correlations, a cross section, and a lithologic column (Plate 2), and a

supplementary map covering part of the Chapter 2 study area (Plate 3). The

accompanying text provides a geologic history to the region, detailed descriptions of

Quaternary units, a discussion of the structural geology along the two graben-bounding

fault zones, and a brief overview of geologic hazards. The text and plates 1 and 2 have

3 been submitted to the Utah Geological Survey and are in preparation for publishing under their Miscellaneous Publication series.

Chapter 5 is a summary of these three studies, which includes their practical and scientific implications.

4 CHAPTER 2

QUATERNARY SALT TECTONICS AND TRANSIENT INCISION OF THE MOAB-

SPANISH VALLEY REGION RECORDED BY CHRONOSTRATIGRAPHY

OF DEFORMED ALLUVIAL DEPOSITS

ABSTRACT

This research uses a record of fluvial deposits and terraces in and around the

Moab-Spanish Valley salt graben in east-central Utah to investigate patterns of faulting, subsidence, and incision. In the central Colorado Plateau, local Quaternary salt deformation is superimposed on region-wide patterns of Colorado River incision. The terrace record of Colorado River tributaries Mill and Pack creeks provides insight into the co-evolution of these processes. Luminescence and cosmogenic burial dating together with topographic surveys demonstrate that a suite of five terraces and undifferentiated alluvial deposits <200 ky old are offset along fault zones and down-warped along the axis of the Moab-Spanish Valley salt graben. The middle-to-late-Pleistocene fault-slip rate at the graben margin is 439 ± 49 m/my. Nearby incision in Mill Creek since ~200 ka has kept pace with graben faulting at 529 ± 44 m/my, which is within range of regional mainstem Colorado River incision rates. These results support the hypothesis that salt- dissolution subsidence of the Moab-Spanish Valley graben is coupled to incision and baselevel fall on the Colorado River via groundwater flow, which is responsible for the below-grade mass removal of dissolved salts. Along the northeast graben rim, the age and distribution of upland gravels suggest subsidence did not begin until after ~1.5 Ma, implying that this structure and the canyon relief of the surrounding landscape are indeed

5 Quaternary features. Tributary incision rates ~15 km upstream from the graben

accelerated from <50 m/my to >600 m/my in the middle Pleistocene. This wave of

incision passing through the drainage network, in the absence of a local source of

differential uplift, originated due to downstream baselevel fall along the Colorado River

and transferred across the Moab-Spanish Valley graben. These results add to a growing

body of evidence that youthful, transient incision is a hallmark of the Colorado River

system, but the source of baselevel fall is poorly understood.

INTRODUCTION

Fluvial terraces record the past position of river channels and provide a datum

from which to measure deformation and incision. In the Colorado Plateau, recent studies

have utilized terraces to infer the timing and patterns of late Cenozoic river incision and

epeirogenic uplift (e.g. Pederson et al., 2006; Cook et al., 2009; Darling et al., 2012;

Pederson et al., 2013a). These regional incision patterns are overprinted by more

localized deformation in the ancestral Paradox Basin of the central Colorado Plateau,

where salt diapirism and dissolution-induced subsidence are active in the Quaternary (e.g.

Colman, 1983; Harden et al., 1985; Jochems and Pederson, 2015). Accordingly, the

distribution, geometry, form, and age of abandoned fluvial terraces in this region should

record a history of salt tectonism and its interaction with larger-scale patterns of river incision.

We focus on the Moab-Spanish Valley salt graben near the city of Moab, Utah, which contains Quaternary fluvial deposits along the Colorado River tributaries of Mill and Pack creeks (Fig. 2.1). Previous workers have hypothesized that the subsidence of

6 Moab and Spanish valleys has deformed Quaternary features and is still active today.

Richmond (1962) mapped Quaternary deposits in the La Sal Mountains and Spanish

Valley from air photos, observing that terraces along Pack Creek converge downstream.

Likewise, Harden et al. (1985) used soil correlations to hypothesize that Pack Creek

terrace deposits in upper Spanish Valley are age-equivalent to subsurface deposits

exposed in gravel pits in the central part of the valley, implying subsidence over the last

150 ky. More recently, Guerrero et al. (2015) calculated a surprisingly rapid late-

Holocene slip rate of ~3 mm/yr (3 m/ky and 3000 m/my) from paleoseismic analysis across a fault strand along the northeast margin of Moab Valley.

Despite these indicators of active salt deformation, rates and spatiotemporal

patterns of graben faulting and subsidence remain unconstrained. Existing geologic maps

(Hackman, 1956; Richmond, 1962; Doelling, 2004) are of insufficient resolution to trace

individual terrace levels across the graben-margin deformation zone in Moab and Spanish

valleys, and correlations between terraces and subsurface deposits (Harden et al., 1985)

lack numeric age control from modern geochronology methods. The slip rate from

Guerrero et al. (2015) represents only late-Holocene displacement across a single fault strand within a deformation zone (Doelling et al., 2002; Chapter 4). Furthermore,

quantifying spatially-integrated and longer-term deformation rates is pertinent to the city of Moab in light of the unresolved geologic hazard potential of salt-dissolution faulting

(Huntoon, 1999; McCalpin, 2009; Carbonel et al., 2013; Gutiérrez et al., 2014). Finally, it is unknown how local subsidence relates to regional incision of the Colorado River, and if tributaries record evidence for the wave of incision recognized 50 km upstream on the

Colorado River (Jochems and Pederson, 2015). These fundamental questions about the

Figure 2.1. Location of Moab and Spanish valleys and Colorado River tributaries Mill and Pack creeks. White polygons denote chronostratigraphy focus areas. Red lines are previously-recognized Quaternary faults (U.S. Geological Survey and Utah Geological Survey, 2006). Abbreviations refer to locations discussed in the text: MF = Moab fault, MR = Moab Rim, KHFZ = Kayenta Heights fault zone, JUOT = Johnson’s Up On Top, JR = Johnson’s Ridge, SMP = South Mountain piedmont, WM = Wilson Mesa, SM = South Mesa.

temporal development of the Moab-Spanish Valley salt graben and its relation to region-

wide processes of landscape evolution motivate our work.

The goals of this research are to generate a chronostratigraphic record of terraces in Mill and Pack creeks, to survey and map their deformation, and to calculate rates of salt-dissolution subsidence, fault slip, and stream incision around the Moab-Spanish

Valley salt graben. This work documents graben subsidence contemporaneous with

baselevel fall on the Colorado River, as well as tributary incision rates that accelerate

8 dramatically over the middle to late Pleistocene. These results are consistent with a wave

of incision passing through the tributary system, as has been recognized elsewhere on the

Colorado Plateau.

BACKGROUND

Cenozoic uplift and incision of the Colorado Plateau

The striking canyon topography of the Colorado Plateau is a result of Cenozoic

uplift and erosion. The early-Cenozoic Laramide orogeny resulted in localized uplift, folding, and basin formation. The middle to late Cenozoic has been marked by epeirogenic uplift, the source of which is heavily debated (e.g Hunt, 1956; Morgan and

Swanberg, 1985; Pederson et al., 2002). Hypotheses include lithospheric hydration, heating, and buoyancy-modification related to removal of the Farallon slab after

Laramide time (Humphreys et al., 2003; Roy et al., 2009), asthenosphere convection and mantle drip over the past 10 Ma (Moucha et al., 2009; van Wijk et al., 2010; Levander et al., 2011; Karlstrom et al., 2012), and isostatic rebound from the erosion of the Colorado

River system (Pederson et al., 2002; Lazear et al., 2013).

It is widely accepted that the late-Cenozoic period of canyon cutting began with the integration of the Colorado River system through the Grand Canyon. Sediment from the Colorado Plateau appears abruptly in the stratigraphic record above the Hualapai

Limestone in the Basin and Range Province sometime between 6.0 and 4.4 Ma (Spencer et al., 2001; Faulds et al., 2001; Howard et al., 2015), which represents the time at which the Colorado River connected with a lower baselevel off of the Colorado Plateau. The ~2 km of Cenozoic uplift provided the potential for deep fluvial incision after drainage

9 integration, which set off an upstream-propagating wave of incision (Pederson et al.,

2002; Dorsey et al., 2007). This period of Pliocene-Pleistocene river incision is

responsible for much of the regional exhumation, as corroborated by thermochonologic

data indicating a pointed acceleration in exhumation rates beginning around 5 Ma in the

central Colorado Plateau (Lee et al., 2013; Murray et al., 2016). Whether this pulse of

incision is still migrating through the Colorado River system is incompletely understood,

but several workers have documented evidence for transient incision in the central

Colorado Plateau (e.g. Cook et al., 2009; Darling et al., 2012; Pederson et al., 2013a;

Jochems and Pederson, 2015).

In Glen Canyon, longitudinal profile analysis and cosmogenic nuclide dating of

terraces imply an acceleration of incision occurred sometime after 500 ka, with rates

increasing from ~100 m/my (0.1 m/ky) to greater than 300 m/my (0.3 m/ky) (Cook et al.,

2009; Darling et al., 2012). These middle-to-late-Pleistocene incision rates in Glen

Canyon are greater than those downstream in the Grand Canyon of ~140 m/my (0.14

m/ky) (Pederson et al., 2006), which is consistent with the recent passing of a wave of

incision. Cook et al. (2009) also modeled incision through upstream-dipping lithology analogous to upper Marble Canyon and lower Glen Canyon. They suggest that the large

knickpoint in the Colorado River at Lee’s Ferry represents a transient incision signal stalled on resistant bedrock, but that an upstream-migrating component of this incision

may have made its way through Glen Canyon in the Pleistocene.

Jochems and Pederson (2015) document unexpectedly rapid late-Pleistocene incision rates increasing upstream from 600 m/my (0.6 m/ky) to 900 m/my (0.9 m/ky) along a reach of the Colorado River from Castle Creek to Dewey Bridge (30-50 km

10 above Moab). These rates do not match the hypothesized “bullseye” of rapid incision in

the central Colorado Plateau (Pederson et al., 2013a), which predicts incision rates of

300-400 m/my (0.3-0.4 m/ky) around Moab if patterns of erosion follow only broad

isostatic rebound. Instead, Jochems and Pederson (2015) suggest the high incision rates at

Dewey Bridge record transient incision, which may be currently stalled in the Westwater

Canyon knickzone farther upstream. It is unknown, however, whether such patterns are

necessarily related to the integration of the Colorado River, or if there is a yet-

unrecognized source of more recent baselevel fall.

Salt deformation in the central Colorado Plateau

Localized deformation of evaporites from the ancestral Paradox Basin is

superimposed on the broader uplift and incision history of the central Colorado Plateau.

The Paradox Basin formed in the middle Pennsylvanian due to flexural-subsidence on the southwest margin of the basement-cored Uncompahgre Uplift, and extends across a northwest-trending belt in the four corners region that is roughly 325 km long and 150 km wide (Fig. 2.2) (Stevenson and Baars, 1986; Barbeau, 2003). Through the late

Pennsylvanian up to 2500 m of evaporites precipitated from hypersaline brines, which included 33 cycles of anhydrite, dolomite, organic-rich calcareous shale, and halite (Raup and Hite, 1992; Trudgill, 2011). Paradox Formation salts began to passively deform shortly after deposition due to differential pressure of overlying sediments. The low- density salts migrated laterally to weaknesses along preexisting faults where they upwelled as diapirs and as salt walls up to 4500 m high, deforming the overlying strata into a series of northwest-trending folds (Trudgill, 2011). Salt migration diminished after

11 the Triassic and was localized along anticline crests, before later burial from Mesozoic

sedimentation (Doelling et al., 1988). Laramide shortening then generated a fold and fault

belt that reoccupied the trend of Triassic salt structures and Pennsylvanian faults

(Doelling et al., 1988).

Figure 2.2. Regional map of the ancestral Paradox Basin. Salt graben are marked by white dashed lines, major folds are shown in black from Doelling (2004), and extent of evaporate facies according to Nuccio and Condon (1996) is marked by the black dashed line. Red box shows location of Figure 2.1.

Regional salt structures were exhumed with the middle-to-late-Cenozoic uplift and erosion of the Colorado Plateau. It is hypothesized that as canyon incision lowered baselevel and groundwater tables, meteoric waters began to dissolve and transport away

12 the Paradox salts, leading to the localized collapse of overlying strata (Cater, 1970;

Doelling et al., 1988). Dissolution has been pervasive in anticline cores where evaporites

are closest to the surface, producing the suite of modern-day, northwest-trending, salt-

collapse graben that occupy the crests of breached anticlines. The cores of these breached

salt anticlines are today capped by insoluble gypsiferous residue from the Paradox

Formation, and this caprock is exposed along the margins of Moab Valley and other salt

graben (Baars and Doelling, 1987; Doelling et al., 2002).

Salt-dissolution subsidence and localized diapirism are believed to be still active in places across the region (Colman, 1983; Doelling et al., 2002; Gutiérrez, 2004; Furuya et al., 2007; Guerrero et al., 2015). The Needles fault zone is a gravity-driven, lateral- spread, extensional belt atop the Paradox Formation east of Cataract Canyon (Huntoon,

1982). Estimates for the initiation of faulting and spreading range from 1.0-0.2 Ma

(Geiger, 2014) to less than 100 ka (Biggar and Adams, 1987), and InSAR analysis shows this area to be actively deforming at maximum rates of 2-3 mm/yr laterally (Furuya et al.,

2007). Chronostratigraphy in graben of the Needles fault zone shows that Holocene drainage and sedimentation patterns have been influenced by local extension and collapse

(Geiger, 2014). In another example, upwelling of the Onion Creek salt diapir 30 km

northeast of Moab may have begun 3-2 Ma based on a unique history of drainage capture

(Colman, 1983). Other studies of ongoing salt subsidence rely on stratigraphic relations and relative ages of Quaternary deposits along the floors of salt graben (e.g. Harden et al.,

1985). Despite these efforts, the rates and timing of active salt deformation throughout the region remain poorly constrained and imprecise due to a lack of numeric dating.

13 Evidence for the deformation of late-Pleistocene alluvial deposits has also been documented by Jochems and Pederson (2015) around the Salt-Cache Valley graben, which is 30 km upstream on the Colorado River from Moab. Here, three levels of

Colorado River terrace deposits with ages between 60 and 25 ka thicken through

Professor Valley, with the bedrock strath of the lowest terrace (M2) plunging below grade. This subsidence in Professor Valley appears to be countered by uplift along the salt wall of the Salt-Cache Valley graben shortly downstream, where straths may be back-tilted upstream and basal gravels from a ~50 ka terrace (M4) are deformed by localized sinkhole collapse (Jochems and Pederson, 2015). Using these relations,

Jochems and Pederson (2015) calculate a late-Pleistocene subsidence rate of ~500 m/my in Professor Valley.

Geologic setting and previous Quaternary research in the Moab-Spanish valley area

The Moab-Spanish Valley salt graben is northwest-trending, ~25 km long, 2.5 to

5 km wide, and bound by fault zones along its southwest and northeast margins. The graben is comprised of two alluvial valleys, Spanish Valley to the southeast and Moab

Valley to the northwest (Fig. 2.1). These depocenters are separated by a bedrock saddle

with thin Pleistocene gravel cover. The graben was formed by the collapse of the Moab

Anticline, which is still preserved to the northwest into Arches National Park as a rollover anticline in the hangingwall of the Moab fault (McKnight, 1940; Doelling et al.,

1988, 2002). The Moab fault is the only tectonic (non-salt) fault in the study area, and is either concealed beneath Quaternary alluvial deposits along the axis of the Moab-Spanish

Valley graben (Doelling et al., 2002; Doelling, 2004) or forms the graben-margin

14 deformation zone along the southwest escarpment (McKnight, 1940; Black et al., 2004).

Further discussion of the structural geology and neotectonics of the Moab-Spanish Valley

graben is included in Chapter 4.

The Colorado River crosses the graben transversely at its northwest end. Along its

5 km reach through the graben it is unconfined and flows across a low-lying floodplain marsh, whereas both upstream and downstream the river is entrenched in ~200 m deep bedrock canyons. To the east of Moab and Spanish valleys lie the La Sal Mountains, which are an exhumed Oligocene laccolith of porphyritic trachyte composition (Nelson et al., 1992; Doelling, 2004). These mountains are the headwaters for two Colorado River tributaries that enter and cross the Moab-Spanish Valley graben (Fig. 2.1). Mill Creek is the northern tributary, with two forks that converge just outside of the northeast margin of the graben. It enters perpendicularly into the graben in eastern Moab Valley, crossing the Kayenta Heights fault zone (Doelling et al., 2002; Chapter 4). Pack Creek enters the graben axially from the southeast, and flows down the length of Spanish Valley before crossing the bedrock saddle into Moab Valley and converging with Mill Creek 2.5 km upstream of the Colorado River. Elevations within these high-relief watersheds range from 3850 m ASL at the summit of Mount Mellenthin to 1200 m ASL along the

Colorado River in Moab Valley.

Early surficial mapping and investigations of Quaternary deposits in the study area were completed by Richmond, (1962), who inferred correlations of gravel deposits around Spanish Valley to glacial cycles extending to the early Pleistocene. Richmond observed that terraces along Pack Creek converge in upper Spanish Valley, are absent in central Spanish Valley, and reappear near the bedrock saddle with Moab Valley. Harden

15 et al. (1985) estimated that these terrace deposits date to the Bull Lake (Marine Isotope

Stage 6) glaciation based on soil profile analysis. Buried soils with similar development

were found exposed in subsurface gravel pits in central Spanish Valley and inferred to be

correlative with the terraces upstream. These observations prompted the hypothesis that

central Spanish Valley had subsided in the last 150 ky. Harden et al. (1985) also reasoned

that initial collapse of the Moab anticline must have post-dated the deposition of high

gravels of Johnson’s Ridge, which are perched on a narrow ridge between South Mill

Creek and Spanish Valley (Fig. 2.1) and inferred to be early to middle Pleistocene in age

(Harden et al., 1985).

More recent mapping and research in the area suggests that salt-dissolution subsidence is an active process in modern times. Doelling et al. (2002) mapped the Moab

7.5’ Quadrangle, and documented that multiple terrace levels along Mill Creek abruptly terminate as the drainage enters the graben in eastern Moab Valley. The absence of terraces along Mill and Pack creeks and the Colorado River through Moab Valley suggests that sediment is being deposited into an actively-subsiding basin (Doelling et al.,

2002), and that subsidence matches or exceeds incision along the Colorado River

(Guerrero et al., 2015). Guerrero et al. (2015) conducted a paleoseismic study across a

strand of the Kayenta Heights fault zone along the northeast edge of Moab Valley, calculating an unexpectedly high late-Holocene average slip rate of ~3 mm/yr. This is

over twice the slip rate of the heavily-studied Wasatch fault zone. With an apparent average displacement recurrence interval of ~316 years (Guerrero et al., 2015), these data portray very active deformation along the graben margin in response to subsurface salt dissolution. Nonetheless, these rates apply to a single strand in a wide and heterogeneous

16 fault system (Doelling et al., 2002; Chapter 4) and span only part of the Holocene.

Longer-term, spatially-integrated subsidence rates are needed to better constrain deformation.

Fluvial terraces as geomorphic markers and their application to incision of the Colorado Plateau

Fluvial terraces are planar valley landforms cored by alluvium that represent abandoned river channels and floodplains (Pazzaglia, 2013). Terraces are formed when a river incises and the past floodplain becomes stranded above grade. A terrace tread is the upper landform surface, the strath is the erosional contact below alluvium. Terraces are commonly classified into two morphologic categories. Strath terraces consist of a thin layer of alluvium (equal to or less than the channel scour depth) overlying the basal strath and represent periods of lateral planation and bedload preservation along a river

(Pazzaglia, 2013). Fill terraces may be underlain by tens of meters of alluvium, often have irregular basal contacts representing valley-bottom paleotopography, and are formed during periods of aggradation (Pazzaglia, 2013). Whether rivers form strath or fill terraces depends on the balance between stream power (the product of discharge and channel gradient) and sediment load and caliber (Bull, 1991). Other important variables

in this distinction include substrate resistance in bedrock streams and valley accommodation space (e.g. Hancock and Anderson, 2002; Wegmann and Pazzaglia,

2002).

Because terrace formation is driven by patterns of river planation, aggradation, and incision, their age is often coupled to hydrologic and climatic cycles. Previous work in the Rocky Mountain region suggests that lateral planation (strath terraces) and vertical

17 aggradation (fill terraces) near headwaters occurs during full glacial conditions, and that

incision and floodplain abandonment occurs during glacial-to-interglacial transitions (e.g.

Ritter, 1967; Sinnock, 1981). Modeling agrees with this relationship, but shows that the

depositional and erosional responses governing terrace formation can lag behind climate

perturbations by thousands of years (Hancock and Anderson, 2002). Terraces along the

major rivers of Colorado Plateau display temporal correlations across the drainage

system. Distinct levels of strath and fill terraces at multiple sites on the Green and

Colorado rivers have been dated to ~15-25, ~35-50, ~60-70, and ~85-100 ka (Aslan and

Hanson, 2009; Pederson et al., 2013b; Jochems and Pederson, 2015). These ages differ from the aggradational episodes predicted by the sequence of Rocky Mountain glaciations. An additional complication is that terrace records in dryland tributary catchments may not be synchronous with mainstem rivers, as aggradation and incision in these smaller systems is likely driven by sediment supply from weathering-limited hillslopes rather than hydro-climatic signals from distant headwaters (Anders et al.,

2005). Specifically, Anders et al. (2005) document tributary fill terrace depositional ages in eastern Grand Canyon that coincide with periods when the Colorado River was incising, suggesting distinct, local tributary responses to climate change.

Fluvial terraces can be used as markers for measuring landscape change. Common applications include using terraces to measure tectonic deformation and calculating river incision rates due to baselevel fall (e.g. Rockwell et al., 1984; Merritts et al., 1994). This approach has been used in the Colorado Plateau as evidence for an upstream-propagating wave of incision related at least partly to the integration of the Colorado River through the Grand Canyon (e.g. Cook et al., 2009; Darling et al., 2012; Jochems and Pederson,

18 2015). In transient incision, rates are expected to be highest at the front of advancing

incision and lower in both the upstream and downstream directions, where the river is

still buffered from the baselevel fall and where it is more adjusted to the new baselevel,

respectively. The interpretation of terrace ages is complicated in the presence of a

transient knickpoint, which can leave downstream terraces in its wake that are

progressively younger upstream, yet mappable as continuous deposits (Pazzaglia, 2013).

Geochronology

Geochronology is fundamental to the use of terraces as geomorphic markers.

Many modern studies of Pleistocene and Holocene alluvial deposits employ optically stimulated luminescence (OSL) or infrared stimulated luminescence (IRSL) dating, which provide an age estimate of the time since sediment was last exposed to sunlight

(Huntley et al., 1985). If the sediment was adequately exposed to light during transport,

the luminescence age gives the time of deposition and burial. Once buried, the

luminescence signal accumulates by the exposure of quartz and feldspar grains to

radiation sources. When mineral grains are once again exposed to light in a laboratory –

blue-green (~470 nm) for OSL and infrared (~880 nm) for IRSL – the electrons return to

their lower energy state and emit a luminescence signal that can be measured by a

photomultiplier tube (Rhodes, 2011). The luminescence signal is a function of the

equivalent dose – the amount of radiation the grain has been exposed to – which is in turn

a function of the environmental dose rate and the time since last light exposure (Rhodes,

2011). Incomplete exposure to sunlight during transport and deposition (partial

bleaching) is a concern for OSL dating of Holocene fluvial sediments, though older

19 Pleistocene sediments are less susceptible to this issue because the residual dose is small

compared with the age of the sample (Murray and Olley, 2002; Rittenour, 2008). OSL

can date quartz sand and silt grains back to ~200 ka (Rhodes, 2011), while IRSL is

commonly limited to feldspar but has a datable range that extends to ~600 ka (Hütt et al.,

1988).

Cosmogenic radionuclide (CRN) isochron burial dating is another technique for

measuring the time since sediment was buried, and it is applicable for deposits that are

older than ~300 ka. This method has a maximum datable range of ~5 Ma (Granger, 2013)

and has been used to date fluvial deposits as old as 4.5 Ma in the Colorado River system

(Matmon et al., 2012). CRN isochron burial dating relies on known cosmic-ray

production and decay rates of 26Al and 10Be in quartz (Granger, 2006). Quartz that is buried under several meters of sediment is sufficiently shielded from further nuclide production such that 26Al and 10Be will decay according to their half-lives. By measuring the 26Al/10Be ratio with accelerator mass spectrometry on multiple samples from the same

position in the deposit, a best-fit isochron can be drawn. The slope of the isochron relates

to the burial age, while the y-intercept records post-burial nuclide production (Balco and

Rovey, 2008; Granger, 2013). The isochron burial method requires samples that have

variable inheritances (different nuclide ratios prior to burial) to allow for enough spread

to calculate a statistically-meaningful isochron.

20 METHODS

Mapping

The distribution and geometry of bedrock, Quaternary deposits, and geologic structures were mapped on 1:8000 scale ortho-rectified aerial imagery. The map area is a

7.5’ quadrangle-sized rectangle. It includes the floor and flanks of Spanish Valley, Pack

Creek from the Moab-Spanish Valley saddle to the confluence with Brumley Creek, and both forks of Mill Creek from Moab Valley to Wilson and South mesas (Fig. 2.1). Field lines were digitized in VrTwo, a 3D photogrammetry and mapping program, and exported to ArcGIS for the final map layout at 1:24,000 scale (Chapter 4, Plate 1).

Surveying

Topographic surveys are necessary to document any deformation of terraces and calculate subsidence and incision rates. Four cross-valley transects along both Mill and

Pack creeks were surveyed, in addition to one valley-parallel survey on a fill deposit where Mill Creek enters Moab Valley (Figs. 2.3 and B.1). Survey lines were chosen to capture the maximum number of terrace levels exposed along a given reach, with transects named in consecutive order from downstream to upstream. Lines Mill1-Mill3

and Pack1-Pack3 were surveyed with a real-time kinematic corrected (RTK) TopCon

HiPer GA GPS and lines Mill4, Mill5, and Pack4 were surveyed with a handheld Trimble

GeoXT GPS. The handheld unit was used for survey lines without reasonable road access to establish a RTK base station, and data were post-process differentially-corrected using the MONTICELLOUT2006 CORS base station. Complete survey data, including tables

21 and graphs for each transect, are included in Appendix B (Tables B.1-B.9; Figs. B.2-

B.10).

Figure 2.3. Location of topographic survey transects along Mill and Pack creeks. Transect lines shown in pink, names key to Appendix B. White dots show kilometer markers upstream from the Colorado River for each tributary, and are referenced throughout the text as “RK” for river-kilometer.

Luminescence geochronology

Fifteen sediment samples were collected for luminescence dating. Six were collected along Pack Creek from two terrace levels and one gravel pit, and eight were collected along Mill Creek from five terrace levels and one undifferentiated alluvium package. Sampling targets were sand lenses in gravelly terrace deposits, where sediment was horizontally extracted in steel pipes and representative samples of the surrounding material collected for dose rate and water content. We preferentially sampled sand lenses

22 with primary sedimentary structures and no evidence of bioturbation or soil development

– qualities indicative of transport and deposition after complete exposure to light without significant post-depositional mixing. Where terrace deposits were thin, cobbly, and devoid of sand lenses, bulk samples were collected at night, transported in doubled black plastic bags, and the sandy matrix later extracted in the dark room lab.

Samples were processed and analyzed at the Utah State University Luminescence

Lab according to the single-aliquot regenerative-dose (SAR) protocol of Murray and

Wintle, (2000) for OSL and Wallinga et al., (2000) for IRSL. IRSL analysis on feldspar grains was performed for samples originally estimated to be older than 150 ka, while younger samples were analyzed with OSL. The environmental dose rate was calculated from inductively coupled plasma mass spectrometry (ICP-MS)-measured concentrations of radioactive elements surrounding the sample site using the conversion factors of

(Guérin et al., 2011). Depositional ages were calculated by dividing the equivalent dose by the dose rate and using the central age model of Galbraith and Roberts (2012). All sample ages are reported to two standard error, and are considered preliminary with 2-17 accepted aliquots.

Cosmogenic radionuclide geochronology

Two upland gravel deposits were dated with the CRN isochron burial technique.

Eight cobbles of varying lithology were collected at a single depth near the basal strath from each deposit. The clasts analyzed and used to calculate isochrons are various sandstones and mudstones from the middle and upper Jurassic sections, as we were unable to separate pure quartz from the Oligocene granodiorite porphyry cobbles due to

23 microcrystalline impurities. The Johnson’s Ridge sample was collected from a depth of

17 m in the T7 upland gravel deposit between South Mill Creek and central Spanish

Valley. This site sits 150 m above the active channel of South Mill Creek and 240 m above the floor of Spanish Valley. The South Mountain sample was collected from a depth of 26 m in the T7 upland gravel south of Pack Creek. This deposit is at the edge of a broad piedmont bajada around South Mountain, and the sample site is 150 m above the active channel of Pack Creek.

CRN samples were prepared and processed at Utah State University (USU) and the Center for Accelerator Mass Spectrometry (CAMS) at Lawrence Livermore National

Lab according to the procedures of Ditchburn and Whitehead (1994). The 26Al and 10Be concentrations and uncertainties for each clast were input into the Granger (2013) linearization code to generate an isochron and calculate a burial age. The CRN results we report here are preliminary, with four and three clasts populating the Johnson’s Ridge and

South Mountain isochrons, respectively.

Rate calculations

Incision rates at one site in Pack Creek and two sites in Mill Creek were calculated from chronostratigraphy data using a linear regression of age and strath elevation from three or more terrace levels. The modern channel was not included as a data point because it represents a dynamic reference frame that can bias incision-rate calculations over short time intervals (Gallen et al., 2015). Numerical ages and 1σ uncertainties along with GPS-surveyed strath elevations and instrument vertical precision

24 values were regressed in MATLAB following the equations of York et al. (2004) to yield incision rates with 1σ uncertainties.

A slip rate across the Kayenta Heights fault zone (KHFZ) east of Moab Valley was calculated using age-equivalent deposits spanning the graben-margin deformation zone along Mill Creek. The difference in surveyed height-above-grade of the samples was divided by their average age, provided that the ages overlapped in 1σ error. The GPS vertical precision and 1σ error in deposit age for each data point were incorporated in the slip rate uncertainty by combining equations for arithmetic error propagation:

= + ; = + + , (1) 2 2 2 𝑥𝑥 𝑎𝑎 𝑏𝑏 − 𝑐𝑐 𝜎𝜎𝑥𝑥 �𝜎𝜎𝑎𝑎 𝜎𝜎𝑏𝑏 𝜎𝜎𝑐𝑐 = ; = ( ) + ( ) + ( ) , (2) 𝑥𝑥 𝑎𝑎 𝑏𝑏 𝑐𝑐 𝑎𝑎∗𝑏𝑏 𝜎𝜎 𝜎𝜎 2 𝜎𝜎 2 𝜎𝜎 2 𝑥𝑥 𝑐𝑐 𝑥𝑥 � 𝑎𝑎 𝑏𝑏 𝑐𝑐 where a, b, and c are measured variables (such as GPS elevation or luminescence age) with standard deviations of σa, σb, and σc, x is a value calculated from these variables

(such as fault slip rate), and σx is the standard deviation of that value (Caldwell and

Vahidsafa, 2013). Using these equations, the cumulative uncertainty of the KHFZ slip rate is reported as its standard deviation.

A maximum total subsidence rate was estimated for Moab Valley based on the greatest subsurface thickness of alluvium in the basin penetrated by a borehole (Doelling et al., 2002) and an extrapolated spline surface of the underlying Paradox caprock constrained by water well depths-to-bedrock (Lowe et al., 2007; Chapter 3 Fig. 3.9).

Subsidence in Spanish Valley was quantified by measuring the angle of terrace downwarping along Pack Creek relative to the background stream gradient. Projecting

25 this angle into the center of Spanish Valley and calculating the depth-below-grade of age-

equivalent gravels provided a subsidence estimate.

RESULTS

Overview

Alluvial deposits in the study area are found at seven correlatable levels in Mill

and Pack creeks and as undifferentiated alluvium beneath the floor of Moab and Spanish

valleys. These deposits are generally clast-supported cobble to boulder-cobble gravel that is moderately-sorted, rounded, and imbricated, with a matrix of fine to medium-grained

sand. Clast composition is consistent across terrace levels, with ~75% porphyritic

trachyte from the La Sal Mountains and ~25% sandstone, siltstone, and chert from mid-

catchment upland benches and bedrock canyons. Terrace deposits were distinguished

based on their landscape position, with deposits mapped as Qatgn and landform tops

(treads) referred to at Tn, where n increases with height above the modern channel

(Chapter 4). For simplicity, terrace and deposit levels both will be subsequently denoted by Tn (for example, T3 for Qatg3 deposits). Lower terrace levels (T1 – T5) were

luminescence dated, T6 is undated, and ages for T7 deposits were obtained from

cosmogenic radionuclide (CRN) burial dating.

In total, 15 luminescence samples (Table 2.1) and two CRN samples (Table 2.2)

were analyzed across the six deposit levels and exposures of valley-bottom alluvium.

Complete luminescence data, including ages, dose rate measurements, grain size

proportions, equivalent dose probability distribution functions, and radial plots, are

presented in Appendix A (Tables A.1, A.2, and A.3; Figs. A.1 and A.2). Complete CRN

26 data, including ages, sample site information, 10Be and 26Al counts from mass spectrometry, and isochron diagrams, are also presented in Appendix A (Tables A.4, A.5, and A.6; Figs. A.3 and A.4). Central ages from these analyses are ~6 ka for the T1y, ~16-

14 ka for the T1, ~78-57 ka for the T2, ~82 ka for the T3, ~93 ka for the T4y, ~194 ka for the T5, ~1.73-1.58 Ma for the T7, and ~137-106 ka for the valley-bottom alluvium.

Deposit ages are not always synchronous between Mill and Pack creeks. Terrace deposits thicken and converge across the margins of the Moab-Spanish Valley salt graben. Strath terraces in the bedrock canyons of Mill Creek terminate and some correlate to undifferentiated alluvium exposed across the Kayenta Heights fault zone at the bowling alley site as the creek enters Moab Valley. Meanwhile, fill terraces in upper Pack Creek thicken downstream and converge with the surface of the Pack Creek fan comprising the floor of southeastern Spanish Valley.

Distribution, geometry, and chronostratigraphy of Mill Creek deposits

Survey and geochronology efforts along Mill Creek targeted three reaches: 1) the bedrock canyon of South Mill Creek below Johnson’s Ridge; 2) the reach 0.5-1.5 km upstream of the Moab Valley graben margin; and 3) the exposed undifferentiated

alluvium deposits where the creek enters eastern Moab Valley. These target areas are referred to as the RK 19 reach, the trailhead reach, and the bowling alley site, respectively. Seven deposit levels were documented; all but the two highest (T6 and T7) were found at both the graben margin and in the upstream bedrock canyons (Chapter 4,

Plates 1 and 3). From the mid-drainage bedrock canyon downstream to Moab Valley, general patterns of terrace occurrence are as follows. In the canyons of South Mill Creek

TABLE 2.1. FINAL LUMINESCENCE GEOCHRONOLOGY FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS Sample Latitude Longitude Elevation Depth Dose rate Equivalent dose OD Age ± 1σ Deposit* number (DD)† (DD)† (m) (m) Method§ Aliquots# (Gy/ky)** (Gy )** (%)†† (ka) Mill Creek T1 USU-2505 38.56304 -109.51630 1293 1.9 OSL-b 16 (23) 2.89 ± 0.13 40.06 ± 5.12 22.9 13.84 ± 1.60 T2 USU-2498 38.56169 -109.51772 1304 3.0 OSL-c 20 (26) 2.21 ± 0.10 125.87 ± 17.48 27.6 57.06 ± 6.81 T3 USU-2504 38.56425 -109.51315 1328 2.8 OSL-c 21 (37) 2.69 ± 0.12 217.93 ± 21.82 17.2 81.01 ± 8.81 T3 USU-2578 38.55297 -109.49403 1374 2.4 OSL-c 21 (28) 2.27 ± 0.10 186.80 ± 28.20 30.4 82.43 ± 10.07 T4y USU-2579 38.50890 -109.42657 1611 2.0 IRSL-b 12 (14) 4.66 ± 0.52 304.67 ± 69.62 38.3 93.47 ± 16.77 T5 USU-2506 38.55874 -109.51043 1395 1.9 IRSL-b 12 (15) 4.38 ± 0.48 565.94 ± 46.50 10.1 194.4 ± 27.2 ba, upper USU-2155 38.55885 -109.52442 1297 4.0 OSL-c 18 (28) 2.70 ± 0.12 286.90 ± 25.86 14.8 106.3 ± 11.3 ba, middle USU-1137§§ 38.55862 -109.52481 1283 10.0 OSL-c 13 (26) 1.26 ± 0.07 172.75 ± 22.95 20.3 136.8 ± 16.7 ba, basal USU-2497 38.55646 -109.52388 1284 20.4 IRSL-c 13 (13) 3.63 ± 0.47 375.49 ± 19.16 1.8 132.6 ± 20.6

Pack Creek T1y USU-2255 38.47260 -109.44423 1521 2.5 OSL-c 14 (15) 1.98 ± 0.09 11.79 ± 2.11 28.1 5.96 ± 0.78 T1 USU-2154 38.45943 -109.42851 1595 1.1 OSL-c 14 (16) 2.28 ± 0.11 36.30 ± 2.04 0.0 15.94 ± 1.58 T2 USU-2500 38.44848 -109.37690 1798 2.8 OSL-c 17 (23) 2.46 ± 0.11 163.76 ± 18.98 22.8 66.50 ± 7.45 T2 USU-2501 38.45156 -109.41365 1655 12.6 OSL-c 17 (32) 3.30 ± 0.15 258.47 ± 46.36 33.4 78.22 ± 10.36 gp, upper USU-2254 38.49062 -109.45771 1462 7.8 OSL-c 16 (22) 1.72 ± 0.12 199.67 ± 22.53 16.5 115.9 ± 14.1 gp, lower USU-2251 38.48957 -109.45903 1444 26.4 OSL-c 19 (22) 2.09 ± 0.21 226.15 ± 32.94 28.9 108.2 ± 16.2 *Organized by stratigraphic position; "T" stands for terrace deposit, "ba" for bowling alley site, and "gp" for gravel pit; ba and gp deposits are undifferentiated alluvium. †Coordinates reported in NAD83. §Abbreviations: b = bulk sample, c = core. #Number accepted, total aliquots in parentheses. **Reported with 2σ error. ††Overdispersion (OD) represents variance in equivalent dose beyond measurement uncertainties; included in equivalent dose error. §§ Sample collected by F. Geiger in 2012; final number of accepted aliquots and age updated with rejection criteria from this study.

28 TABLE 2.2. PRELIMINARY COSMOGENIC NUCLIDE GEOCHRONOLOGY FOR T7 DEPOSITS Source Latitude Longitude Elevation Depth elevation Age ± 1σ Sample (DD) (DD) (m) (m) (m) (Ma)* South Mountain 38.43456 -109.39624 1857 26 2700 ± 300 1.73 ± 0.26 Johnson's Ridge 38.48878 -109.41846 1770 17 2300 ± 500 1.58 ± 0.50 *Calculated from isochron using Granger (2013) linearization method.

(which contain better-preserved terraces than North Mill Creek), alluvial gravels <5 m thick form isolated strath terraces that commonly have poorly-preserved treads. The straths of terraces T1-T3 extend above grade from knickpoints, and increase in height stepwise downstream. Before crossing the Kayenta Heights fault zone, Mill Creek flows through a kilometer-long reach near the Mill Creek trailhead that is less confined than reaches immediately above and below (Fig. 2.4). Downstream from the trailhead, Mill

Creek cuts through a short (0.5 km) gorge before emptying into Moab Valley. Terraces

T1-T5 are preserved at the trailhead, but with the exception of T1 they are not found downstream in Moab Valley (Fig. 2.4). Deposits exposed at the graben margin against the

Kayenta Heights fault zone at the Mill Creek bowling alley site are >30 m of

undifferentiated gravel fill. The distribution, geometry, and age of deposits in Mill Creek are reported in detail next in order from lowest (T1) to highest (T7), concluding with

Moab Valley undifferentiated alluvium.

Mill Creek terraces T1-T5

The strath of the Mill Creek T1 emanates from an ~8 m knickpoint at the powerhouse dam near the Mill Creek trailhead at river kilometer (RK) 6.4 (Figs. 2.3 and

2.5). Upstream from this knickpoint for 2 km, T1 gravels are at grade with the bedrock

Figure 2.4. Mapped distribution of alluvial deposits and faults along Mill Creek at the edge of Moab Valley. Tn = Qatgn of Chapter 4 Plate 3, Qagu = undifferentiated alluvium. Green stars denote luminescence sample sites, and are labeled by sample number. River kilometers from Figure 2.3 are shown in white, and red line shows survey transect from Figure 2.6B.

30 channel and part of the modern floodplain. Downstream of the powerhouse dam and upstream of the Kayenta Heights fault zone (RK 5.5), the T1 gravel is 2-3 m thick and 8-

10 m above grade (Fig. 2.6B). An OSL sample in this reach yields an age of 13.84 ±

1.60ka (USU-2505). Immediately upstream of the eastern strand of the Kayenta Heights

fault zone the T1 gravel deposit thins to 1 m atop its basal strath. The deposit is inferred

to thicken once again across the western fault strand where the strath plunges below

grade and the tread is not visibly offset by faulting. Downstream from this point the T1

deposit is at least 4 m thick (Table B.1). Once inside the graben, the T1 tread broadens

and coalesces with the Pack Creek T1 to form the floor of eastern Moab Valley (Fig. 2.4).

The T1 tread eventually converges with the modern channel in central Moab Valley, and

for its final 1.5 km to the Colorado River, Mill Creek flows across a floodplain marsh.

Figure 2.5. Terraces and knickpoint at the Mill Creek trailhead. Straths marked by dotted white lines, geology shown in solid white lines. T1 strath emerges above grade at the powerhouse knickpoint. Photo is taken from T3 tread on the north side of Mill Creek, looking south. Jn = Navajo Sandstone, Jk = Kayenta Formation.

The Mill Creek T2 terrace separates from the active bedrock channel and

floodplain at a knickpoint at RK 8.8. There are discontinuous, ~2 m thick, carbonate-

cemented T2 deposits downstream to the trailhead reach. At the trailhead, there are 1-4 m

Figure 2.6. Mill Creek valley cross sections. (A) South Mill Creek at RK 19 (Fig. B.6; Table B.5). (B) Mill Creek trailhead at RK 6.3 (Fig. B.4; Table B.3). A mix of T1-T3 age gravels are assumed to compose the canyon bottom alluvium in upper South Mill Creek. Views are downstream. Location of cross section lines shown in Figures 2.3 and 2.4.

32 thick T2 deposits on both sides of Mill Creek that are 20-27 m above grade (Fig. 2.6B).

The north-side T2 has a younger fill-cut tread inset 3m below the upper tread surface. An

OSL sample from 1 m above the strath of the south-side T2 yields a depositional age of57.06 ± 6.81 ka (USU-2498). The farthest-downstream T2 is between the east and west strands of the Kayenta Heights fault zone and has a lower strath height of 17 m, potentially representing fault offset.

The Mill Creek T3 exhibits a similar pattern as the T2. The strath departs from the channel around RK 15 in South Mill Creek across a series of bedrock knickpoints that are more subdued than those marking the start of T1 and T2. Upstream from this point to the edge of the study area at RK 26, cobble-gravel deposits that are 1-2 m thick form the

alluvial floor of South Mill Creek Canyon, in places buried by sandy Holocene deposits

(Fig. 2.6A). These canyon-floor gravels are not incised and likely represent deposition

spanning T3 to T1 time across several cut and fill episodes. Downstream of RK 15, 1-3 m

thick T3 deposits are sparsely preserved. An OSL sample at RK 9.3, where the T3 strath

is 22-27 m above grade (Fig. B.5; Table B.4), has an age of 82.43 ± 10.07 ka (USU-

2578). On the north side of the Mill Creek trailhead, 3-5 m thick T3 gravels are preserved

as four conical hills on a strath that dips toward the creek, which likely represents the

bedrock margin of the paleo-valley. The higher end of the strath here is 43 m above grade

(Fig. 2.6B). An OSL sample from the Mill Creek trailhead T3 yields an age of 81.01 ±

8.81 ka (USU-2504), which is consistent with the age of the T3 deposit 3 km upstream.

T4 deposits are coarser than T1-T3 deposits throughout the drainage network, and are preserved in South Mill Creek upstream of RK 18. At RK 19, a T4o (older) has a strath 62-65 m above grade and a T4y (younger) has a strath 53-57 m above grade (Fig.

33 2.6A). Both deposits are 3-4 m thick in this location, and the IRSL age of the T4y is

93.47 ± 16.77 ka (USU-2579). The T4 is not preserved between RK 18 and the graben

margin, but a final <2 m thick gravel deposit at the vertical position of the T4 is present

less than 100 m east of the Kayenta Heights fault zone. This inferred T4 has a strath that

is 70 m above grade (Fig. 2.6B). This deposit could not be dated because it is thin, lacks

abundant sandy matrix, and has been partially quarried.

T5 deposits occur only as incompletely-preserved, remnant gravel knobs and thin strath terraces along the canyon walls of Mill Creek and South Mill Creek. The farthest upstream knob is off the south end of Johnson’s Ridge at RK 22.3, 1-2 m thick, and 100 m above the South Mill Creek channel. Another T5 deposit at RK 16 is 107 m above grade and 3-5 m thick (Fig. 2.6A). T5 deposits are not found between this point and the

Mill Creek trailhead reach. Above and to the south of the trailhead, three conical gravel

deposits mark the T5 remnants (Fig. 2.4), and are 2-4 m thick on straths 105-110 m above the modern channel (Fig. 2.6B). The IRSL age of these deposits is 194.4 ± 27.2 ka (USU-

2506).

Mill Creek upland gravels T6 and T7

The T6 gravel caps the east rim of “Johnson’s Up On Top” above the canyon of

South Mill Creek, and is partially concealed by an extensive dune field (Chapter 4, Plate

1). The Johnson’s Up On Top T6 deposits form a linear trend that parallels the course of

South Mill Creek, but are inferred to extend more broadly across the bench beneath the

dune field. Where the tread is preserved and exposed, these deposits are 5-8 m thick on a

strath that is 148 m above grade at RK 19 (Fig. 2.6A) and 135 m above grade at RK 13.

34 In this area, the T6 strath is 20-30 m below the strath of the thicker T7 gravels which

occupy the same drainage divide on Johnson’s Ridge to the southeast (Fig. 2.6A).

Upstream of RK 18, T6 deposits occur as isolated hills that lack a preserved tread, and

are less than 3 m thick. An inferred T6 at RK 23.5 has a strath that is 117 m above the

channel of South Mill Creek. T6 deposits are absent downstream of RK 13.

The Johnson’s Ridge T7 gravel forms prominent linear upland surfaces and knobs

above South Mill Creek. T7 deposits grade to higher piedmonts on the flanks of the La

Sal Mountains (Doelling, 2004), but are distinct in map view in that they are only preserved in a narrow zone closely paralleling South Mill Creek. These deposits also cap disconnected hills above the South Mill canyon between RK 25 and RK 23 (Chapter 4,

Plate 1). The type-section of the Johnson’s Ridge gravel caps the <2 km wide corridor

along the South Mill Creek/Spanish Valley drainage divide between RK 22 and RK 18

(Fig. 2.7). The Johnson’s Ridge strath is 150-175 m above South Mill Creek (Fig. 2.6A)

and 240-265 m above the floor of Spanish Valley. This landform is up to 250 m wide and

extends ~3 km downstream as a continuous ridge 40-55 m above the broader T6 surface

of Johnson’s Up On Top. The 10-25 m thick Johnson’s Ridge T7 fill exhibits ~15 m of

paleotopography across its strath (Fig. 2.6A). There is a stage V calcrete developed along

the well-preserved tread that is prominently exposed at the upstream edge of Johnson’s

Ridge. Clasts were sampled ~1 m above the strath (17 m below the tread) at this location

for cosmogenic radionuclide (CRN) burial dating (Fig. 2.7). This sample yields an age of

1.58 ± 0.50 Ma. The high uncertainty stems from a single outlier clast containing an

unusually high 10Be concentration (Fig. A.3). The large 1000 m elevation window in the

upper catchment for the source of the sample clasts also adds to the uncertainty in the

35 inherited cosmic dose rate (Granger, 2013). Clasts in the CRN sample outcrop are

imbricated to the west-northwest, suggesting transport away from the La Sal Mountains.

The farthest-downstream T7 deposits are on the east side of South Mill Creek between

RK 15 and 12.5. These deposits form two linear, drainage-parallel landforms similar to

Johnson’s Ridge. The upstream (southern) deposit is 8.5-12.5 m thick atop a strath that is

175 m above grade, while the downstream (northern) deposit is 5-8 m thick atop a strath that is 182 m above grade.

Figure 2.7. Johnson’s Ridge CRN sampling location. Google Earth view is to the northeast, looking downstream along South Mill Creek with the Moab-Spanish Valley salt graben to left.

Mill Creek undifferentiated alluvial gravel deposits

A >30 m thick package of exposed undifferentiated alluvial gravels (Qagu) in the

immediate hangingwall of the Kayenta Heights fault zone lies where Mill Creek enters

36 Moab Valley (Fig. 2.4). Deposits are preserved on both sides of Mill Creek, and their

treads project to the upper surfaces of undifferentiated alluvium deposits in southern

Moab Valley south of Pack Creek. The tread of the southern Mill Creek deposit (at the

Moab bowling alley) slopes valley-ward, but in the distal 70 m is subtly back-tilted to the

east (Fig. B.3; Table B.2). At the low point of this back-tilt, the tread is 30 m above the

valley floor T1 surface and 34 m above the active channel (Fig. B.2; Table B.1). The

undifferentiated alluvium has a poorly-known thickness, but may extend down another 40

m below grade based on water well data (Lowe et al., 2007). Excavated exposures of the

bowling alley deposit show a seemingly continuous package of pebble cobble-gravel.

There are no buried soils in the exposure aside from a well-developed stage III carbonate horizon in the upper meter. Gravels are imbricated from west-northwest to south- southwest, suggesting transport from Mill Creek. Sand lenses are much more common in

Qagu deposits than in upstream terrace gravels and are 10-20 m wide and up to 4 m thick.

Three luminescence samples were collected from sandy interbeds in varying stratigraphic positions in the bowling alley Qagu deposit (Fig. 2.4). OSL samples from 4 and 10 m beneath the tread on the Mill Creek (north) side yield ages of 106.3 ± 11.3 ka (USU-

2155) and 136.8 ± 16.7 ka (USU-1137), respectively. A basal IRSL sample from 20.4 m

beneath the tread on the south side of the deposit yields an age of 132.6 ± 20.6 ka (USU-

2497). The slight discordance between the middle and basal ages may suggest the presence of a ~10 m-amplitude cut-and-fill unconformity in the deposit that was formed between ~136 and ~133 ka. Alternatively, the middle and basal ages are within 1σ error of each other and may simply represent a period of rapid, continuous aggradation.

37 Distribution, geometry, and chronostratigraphy of Pack Creek deposits

The T1, T2, T3 terraces, and the T7 upland gravels are distinguishable and

preserved along Pack Creek and in Spanish Valley. Surveying and geochronology field

work was focused between the farthest downstream gravel pit in Spanish Valley at Pack

Creek kilometer (RK) 16.5 and upstream to the Pack Creek/Brumley Creek confluence at

RK 27 (Fig. 2.8). The T1 forms the bottom of Pack Creek’s alluvial valley upstream of

the graben, and it broadens downstream to form the Pack Creek fan in the graben. The T2

and T3 are 15-40 m thick fill terraces in their upstream study area, and they converge with the basin floor downstream in southeast Spanish Valley. There are several piedmont-slope P4 deposits along the Pack Creek valley sides that appear to grade to the

T4 terrace level along the valley axis, but no Pack Creek terraces of these positions are preserved. T5 and T6 terrace levels are completely absent along Pack Creek. The T7 deposit forms a high mesa south of Pack Creek that represents the northwest edge of an extensive piedmont surface on the flank of South Mountain. The Spanish Valley basin floor is mapped as equivalent to T1, and it is underlain by undifferentiated alluvial

gravels (Qagu) 30-95 m thick (Lowe et al., 2007) that are exposed only in gravel pits.

Farther downstream across the bedrock saddle between Spanish and Moab valleys, the T2

and T3 terraces reappear, before they are no longer found in the bottom of Moab Valley.

Pack Creek terraces T1-T3

The lowest and youngest fluvial terrace, the T1, is preserved along Brumley and

Pack creeks where the drainages exit the bedrock confines of the La Sal Mountains (RK

~28). From the Pack-Brumley confluence to Spanish Valley, the T1 forms the valley

Figure 2.8 (opposite). Map of alluvial deposits in Pack Creek and southeast Spanish Valley. T1 sublevels are grouped here for clarity relative to Plate 1 of Chapter 4. Tn = Qatgn, and Pn = Qapgn of Chapter 4. Green stars denote luminescence and CRN sample sites, and are labeled by sample number. River kilometers from Figure 2.3 are shown in white, and cross sections from Figure 2.9 are displayed as red lines.

39 floor and is incised 7 to 12 m by the modern channel. Inset below the T1 1-2 and 3-5 m,

respectively, are T1i (intermediate) and T1y (youngest) terraces (Fig. 2.9A). The strath is

exposed in the channel bottom on Mancos Shale at RK 25.2 for a <300 m reach, but is

otherwise below grade. Here, the T1 is 10 m thick. The Pack Creek channel near the head

of the Pack Creek fan (RK 22) is incised 8 m below the T1 (Fig. B.8; Table B.7). In this

reach, an OSL sample from a T1 cutbank exposure yields an age of 15.94 ± 1.58 ka

(USU-2154). Between RK 22 and 21, T1 converges with the channel and the T1y

expands to become the most recent depositional lobe of the fan (Figs. 2.9B and 2.10). A

gravel pit OSL sample 2.5 m below the T1y tread at RK 19 gives a depositional age of

5.96 ± 0.78 ka (USU-2255). Downstream from this point across Spanish Valley, Pack

Creek is incised less than 2 m below the T1 tread. Along the crest of the Spanish/Moab

valley saddle Pack Creek flows through a ~2 km long bedrock corridor. Due to the

increased confinement, the formerly extensive T1 is reduced to several discontinuous

deposits with treads 3-5 m above the active channel. The T1 bedrock strath is exposed in

the reach, but is never more than 1 m above grade. Once Pack Creek enters Moab Valley,

the T1 expands and coalesces with the Mill Creek T1 to form the valley floor. In this

reach to its confluence with Mill Creek, the Pack Creek channel is incised 3-6 m into the

T1.

The T2 of Pack Creek is preserved as a 20 m thick fill terrace starting where

Brumley Creek exits its confined bedrock canyon at RK ~28. Here the bedrock strath is

<2 m above the channel, and then disappears below grade. OSL sample USU-2500 from this location has an age of 66.50 ± 7.45 ka. From the Pack-Brumley creek confluence the

T2 is not preserved until RK 25.5, where its tread is 35 m above grade and its strath is

Figure 2.9. Pack Creek valley cross sections. (A) Upper Pack Creek at RK 23.5 (Fig. B.9;

Table B.8). (B) Pack Creek fan at RK 20.7 (Fig. B.7; Table B.6). T3 tread decreases in height above grade as it converges downstream with the valley floor. Views are downstream. Location of cross section lines shown in Figures 2.3 and 2.8.

concealed (Fig. B.10; Table B.9). From that point the T2 tread decreases in height downstream. At RK 23.5 it is 16 m above grade (Fig. 2.9A), and a cutbank exposure yields an OSL age of 78.22 ± 10.36 ka (USU-2501). Less than a kilometer downstream at

RK 23, the T2 tread converges into the subsurface. At the topographic saddle between

Spanish and Moab valleys the T2 locally appears as a 1-2 m thick deposit on a Navajo

Sandstone strath 15 m above Pack Creek (Chapter 4, Plate 3). Beyond RK 8 the T2 is absent in Moab Valley.

There are three remnants of T3 terrace deposits along Pack Creek between RK 25 and 20 (Fig. 2.8). Upstream of this reach, several P3 piedmont deposits project to former

Figure 2.10. Oblique photo of Pack Creek fan in southeast Spanish Valley. White dashed arrows mark the edge of the terrace treads, indicating the direction they dip. The T3 tread converges with the T1 basin floor just upstream of a gravel pit. Note the downstream broadening of the T1y Holocene terrace.

positions of the Pack Creek T3. The tread of the farthest-upstream T3 south of Pack

Creek slopes down-valley from RK 25 to RK 22.4, where the deposit is erosionally truncated and exposes a strath 8 m above the valley floor. The deposit is 17 m thick above a Mancos Shale strath at its upstream end and 34 m thick above a Navajo

Sandstone strath with a tread 50 m above grade at the RK 23.5 cross section (Fig. 2.9A).

Straths exposed in gullies are seen dipping north towards Pack Creek, but are otherwise concealed by colluvium. The second T3 deposit extends from RK 23 to RK 21.8 approximately 1 km south of the Pack Creek channel (Fig. 2.8). The tread of this T3 is 50 m above the valley floor at the upstream end and 25-30 m in height at its downstream erosional truncation, which exposes a Navajo Sandstone strath 10 m above the valley floor. The strath is locally-exposed in south-facing gullies, and the overlying gravels are observed to thicken downstream. The farthest downstream T3 is just north of Pack Creek

42 and occupies a prominent position at the head of the Pack Creek fan, extending from RK

22.2 to RK 20 (Fig. 2.8). The tread is 30 m above Pack Creek at its upper end (Fig. B.8;

Table B.7) and is split into two treads 27 m and 11-14 m above grade at RK 20.7 (Fig.

2.9B). At RK 20, the T3 tread smoothly projects into the subsurface and does not appear

to be erosionally truncated (Fig. 2.10). This T3 deposit does not have a planar basal

strath, as evidenced by a complete gravel exposure on its upper south side. However,

there are Burro Canyon Formation outcrops in at least two locations on its northeast riser

(Chapter 4, Plate 1), which likely record paleo-valley topography. An OSL sample from

this terrace returned an anomalously young age due to capturing a more recent Pack

Creek channel-fill, therefore there is no constraint on the T3. Collectively, the T3

deposits thicken downstream and have down-tilted treads that lose 50 m in height over

3.5 km of distance, with much of that loss occurring in the final 0.7 km.

Pack Creek upland gravel T7

There is a thick and extensive T7 deposit forming the mesa south of Pack Creek

that is connected to a broad, bajada-type piedmont gravel blanketing the lower slopes of

South Mountain. This upland gravel is consistently 40-50 m thick from the Pack-Brumley

creek confluence to its abrupt erosional truncation at the head of a tributary amphitheater

at RK 25 (Fig. 2.8), with a tread 175-180 m above the channel of Pack Creek. The strath

overlies the Mancos Shale, and a 1-8 m thick layer of carbonate-cemented gravel comprises a cliff-forming basal conglomerate. The remainder of the deposit is an unconsolidated cobble and boulder-cobble gravel that is imbricated to the west and is capped by a 1 m thick stage V calcrete. An initial CRN burial sample from 26 m below

43 the tread yields an age of 1.73 ± 0.26 Ma (Fig. 2.11). The isochron producing this age is essentially formed by two data points, as two of the three clasts thus far analyzed have

26Al/10Be ratios that nearly overlap (Fig. A.2), which necessitates the additional two

clasts still in the process of analysis at the time of this writing.

Figure 2.11. South Mountain CRN sampling location. Google Earth view is to the east-

southeast from southeast Spanish Valley, looking upstream towards the La Sal Mountains

and the headwaters of Pack Creek.

Pack Creek undifferentiated alluvial gravel deposits

Unlike in Moab Valley, undifferentiated gravel deposits are not naturally exposed

in Spanish Valley and can only be accessed in gravel pits. In the gravel pit at RK 16.5,

two buried soils with stage III carbonate horizons separate three gravel packages (Fig.

2.12). OSL samples were collected from the lower and upper packages at depths below the surface of 26 and 8 m, respectively. The lower sample has an age of 108.2 ± 16.2 ka

44 (USU-2251) and the upper sample has an age of 115.9 ± 14.1 ka (USU-2254). These results are apparently inverted given their stratigraphic position, suggesting that an obscured, large cut-and-fill contact went undetected in the gravel pit or that the ages are essentially synchronous given their overlapping 1σ error. It is also noteworthy that this apparent age reversal is resolved if a lower in-situ water-content of 5% is used instead of

15% for USU-2254, which is closer to the measured content of 3.1% and implies that the

USU-2254 sample site has been mostly above the water table since deposition (Table

A.2). Regardless, the Spanish Valley gravel pit samples are within the age range of the

Qagu deposits in Moab Valley and fall between the Mill Creek T5 and T4 ages. The buried soils observed in gravel pits are apparently rough constraints on depositional episodes and do not represent major hiatuses. Furthermore, it is likely that these gravel packages and buried soils are laterally complex, as is also apparent with the Moab Valley

Qagu deposits.

Incision and deformation rates

Pleistocene incision rates calculated from the chronostratigraphy of Mill and Pack

creeks vary over distance and time. At the Mill Creek trailhead reach a plot of vertical

channel position through time, as constrained by GPS surveying and luminescence ages of T1-T5, yields an incision rate of 529 ± 44 m/my (0.529 ± 0.044 m/ky) over the past

~200 ky (Fig. 2.13B). In the upper South Mill Creek reach at RK 19, an incision rate of

619 ± 80 m/my (0.619 ± 0.080 m/ky) to 1277 ± 387 m/my (1.277 ± 0.387 m/ky) over the last ~200 ky was calculated from the T1-T3, T4y, and projected T5 terraces (Fig. 2.13A).

These endmember rates reflect uncertainty in the age of the canyon-bottom alluvium in

Figure 2.12. Gravel pit exposure of Spanish Valley upper undifferentiated gravels. The basal ~25 m correlates in age with upper Qagu

deposits in Moab Valley. Overburden may hide a cut-and-fill unconformity responsible for apparent stratigraphic inversion of ages. Upper 3 m of gravel is Holocene in age. Location shown conceptually in Figure 2.16.

46 upper South Mill Creek, which is laterally continuous with downstream T3 deposits but

may also contain T1-age gravels in cut-and-fill packages. The slower rate is similar to the

Mill Creek trailhead rate, and implies steady late-Pleistocene incision, while the faster rate entails rapid incision between T5 and T3 that has been essentially zero since ~82 ka.

Prior to ~200 ka, incision between the T7 Johnson’s Ridge gravel and the undated T6 gravel of Johnsons Up On Top was much slower at <50 m/my (<0.05 m/ky) (Fig. 2.13A).

Figure 2.13. Mill Creek incision history. Plots show depositional age (constrained by luminescence and CRN dating) on x-axis and elevation above grade (measured in GPS surveys) on the y-axis. (A) Incision history in upper South Mill Creek (RK 19). (B) Incision history at the Mill Creek trailhead (RK 6.3). All rates are given in m/my, with 1000 m/my equal to 1 m/ky and 1 mm/yr. Horizontal bars show age 1σ uncertainties. Incision rates are reported with 1σ uncertainties from linear regression in MATLAB using the equations of York et al. (2004).

47 In addition to incision in upstream canyon reaches, this study provides constraints

on faulting and subsidence at the graben margin by correlating terrace gravels in the

trailhead reach of Mill Creek to undifferentiated gravel deposits in Moab Valley (Fig.

2.14). These sites are separated by the Kayenta Heights fault zone (KHFZ), which is the

graben-bounding deformation zone on the northeast side of Moab Valley (Doelling et al.,

2002). The fault zone has two mapped strands where Mill Creek enters Moab Valley

(Fig. 2.4). The eastern strand down-drops Navajo Sandstone against Wingate Sandstone south of Mill Creek, accommodating 100-200 m of throw. This offset rapidly decreases toward the fault tip <1 km north of Mill Creek (Fig. 2.4; Chapter 4, Plate 3). The western strand juxtaposes Qagu deposits against Navajo Sandstone, accommodating at least 70 m of displacement based on water-well depths to bedrock in the basin (Lowe et al., 2007).

Faults of the KHFZ are near-vertical and form relatively short segments averaging ~3 km

in length, which suggests they are shallowly-rooted in the Paradox Formation and respond to subsurface salt dissolution by vertical collapse (Chapter 4). There is confirmed late-Holocene slip on a strand of the KHFZ 3.5 km to the southwest based on paleoseismic analysis (Guerrero et al., 2015), but there is no apparent offset of the ~14 ka

T1 terrace across the KHFZ at Mill Creek. It is possible that subsidence is accommodated

by additional unmapped strands of the KHFZ that are farther southwest in the basin interior. The bowling alley Qagu deposit south of Mill Creek is subtly deformed with a tread that is back-tilted ~0.7° to the east in its distal 70 m (Fig. B.3; Table B.2), which may support the existence of an unmapped fault at the deposit’s southwest edge.

However, there are no documented fault scarps in the alluvium of Moab Valley, so the presence of additional strands of the KHFZ remains speculative.

Figure 2.14. Mill Creek longitudinal profile, incision, and deformation diagram. Moab Valley, the graben margin, and the trailhead reach are shown at left, and the upper South Mill Creek reach is shown to the right (note break in horizontal axis). Numerical dates and rates reported with 1σ uncertainty, and are discussed in the text. Pink dots show age-equivalent deposits used to calculate KHFZ slip rate. Tn = terrace, Qagu = undifferentiated alluvium. Geology shown conceptually: Jn = Navajo Sandstone, Jk = Kayenta Formation, Jw = Wingate Sandstone, TRc = Chinle Formation, TRm = Moenkopi Formation, Pc = Cutler Group.

49 A displacement rate across the two mapped strands of the KHFZ was calculated

from the chronostratigraphy at Mill Creek. The uppermost bowling alley Qagu

luminescence sample (USU-2155) has a depositional age of 106.3 ± 11.3 ka. The T4

strath terrace in the immediate KHFZ footwall is undated, but is estimated to be age-

equivalent to the upper bowling alley gravel package based on age control of the T4y 12

km upstream in South Mill Creek (93.47 ± 16.77 ka). Taking the difference in local heights relative to modern grade between the T4 in the bedrock footwall and the upper hangingwall Qagu sample yields a total displacement of 47 ± 1.4 m, with the 1σ

uncertainty reflecting GPS vertical precision calculated according to equation (1).

Dividing this displacement by the 106.3 ± 11.3 ka age of the upper Qagu sample using

equation (2) yields a displacement rate of 439 ± 49 m/my (0.439 ± 0.049 m/ky).

Net subsidence in the middle of Moab Valley is greater than fault slip on the

KHFZ, as the middle of the basin is much deeper than the margins near the fault zone. A

water well ~300 m northwest of the Mill Creek bowling alley Qagu deposit indicates that

the alluvium is 37.5 m thick (Lowe et al., 2007). Meanwhile, in lower Moab Valley just northwest of the Colorado River the Atlas Minerals ATP-1 borehole penetrated 124 m of

alluvium without reaching bedrock, indicating that the basin is at least that deep

(Doelling et al., 2002). A spline of the bedrock floor of Moab Valley projected in GIS using known alluvium thicknesses from water wells estimates a maximum alluvium depth of 190 m directly beneath the Colorado River (Lowe et al., 2007; Chapter 3). The accumulation of this alluvium may be partly attributed to scouring by the Colorado River during flood events (Weisheit and Fields, 2006), and thus it is necessary to subtract the river’s potential scour depth to estimate total subsidence. Sand-bed rivers may scour to

50 five times the mean channel depth (Best and Ashworth, 1997), which suggests the

Colorado River may scour to ~25 m given a 5 m channel depth. Thus, the projected 190

m of alluvium beneath the Colorado River may be accounted from by ~165 m of

subsidence.

Comparing this estimated subsidence depth to the height-above-grade of the T5 terrace at the Mill Creek trailhead, and assuming the deepest alluvium in Moab Valley is

T5-age, gives a maximum Moab Valley subsidence rate of ~1,350 m/my (1.35 m/ky) over the middle to late Pleistocene. Alluvium at the bottom of Moab Valley is unlikely to be younger than this assumed age since T5-equivalent gravels are exposed at the surface at the Mill Creek bowling alley site. However, it is possible that the deepest alluvium in

Moab Valley is older than T5-age, which would produce a lower total subsidence rate and necessitates the emphasis that ~1,350 m/my (1.35 m/ky) is only a maximum

constraint. Performing an identical calculation using the 125 m of alluvium found in the

Atlas Minerals ATP-1 borehole (Doelling et al., 2002) as a minimum estimate for total

alluvium thickness yields a total subsidence rate of ~1000 m/my (1 m/ky). Repeating

these same equations with the assumption that the basal Moab Valley gravels are much

older at T7-age yields a substantially slower net subsidence rate of <200 m/my (<0.2

m/ky). However, using the T5-age to estimate these rates is most meaningful because it

allows comparison to the slip rate across the KHFZ, which is calculated over a similar

time period. Therefore, the total subsidence rate, though only constrained by maximum

estimates, may be over double the estimated late-Pleistocene slip rate of the KHFZ. If correct, this suggests that greater subsidence in the center of Moab Valley must be

51 accommodated either by additional unmapped salt-dissolution faults or a progressive sagging into the core of the graben.

Incision along Pack Creek has fewer age constraints than Mill Creek and terraces appear to be gradually influenced by subsidence approaching the graben, even though no faults are crossed. An incision rate was calculated near the Pack-Brumley creek confluence upstream from the wide deformation zone where there are ages for the T1, T2, and T7 deposits. This location displays an unsteady incision history marked by aggradation around T7 time followed by rapid incision in the late Pleistocene (Fig. 2.15).

Using surveyed T3-T1 tread heights, local T2 and T1 ages, and an estimated T3 age of

100 ± 20 ka based on correlative Qagu gravel pit samples, late-Pleistocene incision in

Pack Creek is regressed at 501 ± 65 m/my (0.501 ± 0.65m/ky).

Figure 2.15. Incision history at the Pack Creek/Brumley Creek confluence (RK 26.8). Plot axes are the same as in Figure 2.13. Horizontal bars show age 1σ uncertainties. Incision rate is reported with 1σ uncertainty from linear regression in MATLAB using the equations of York et al. (2004).

For Spanish Valley, subsidence is quantified using the estimated angle at which the Pack Creek T3 terraces project below grade (Fig. 2.16). The T3 tread elevation

52 relative to grade was constrained at four locations: the Pack3 (Fig. 2.9A), Pack2 (Fig.

B.8), and Pack1 (Fig. 2.9B) cross section surveys, as well as the valley-bottom position

where the tread converges with the subsurface at RK 20. A linear regression through the

relative elevations and valley centerline distance of these data points gives a best-fit

down-valley tilt angle of ~0.9° relative to stream gradient. Projecting this angle into the

center of Spanish Valley where alluvium is believed to be the deepest (Lowe et al., 2007)

estimates a depth of ~45 m below grade for the top of T3-equivalent gravels. This value

is conservative in assuming that the T3 terrace down-warps gradually and linearly rather

than rolling over or otherwise steepening. This is a realistic assumption given the

consistent down-valley tilt of the surveyed T3 treads and the fact that the projection ends

up in the middle of the total alluvium depth for this part of Spanish Valley, which is

estimated at ~95 m (Lowe et al., 2007; Chapter 3). The luminescence-sampled Qagu

gravel pit exposure (Fig. 2.12) is <1 km from the deepest point in Spanish Valley where

the tilt angle was projected, and has ages of 108.2 ± 16.2 ka and 115.9 ± 14.1 ka. The

gravel pit geochronology suggests the top of ~100 ka T3-equivalent deposits in central

Spanish Valley are locally only a few meters below grade (the depth of USU-2254) rather

than a few tens of meters based on the linear projection of upstream terrace treads to this

point. This discrepancy likely reflects the complex lateral stratigraphy of these lobe- shaped fan deposits in Spanish Valley, as well as the fact that they should thicken downstream. Regardless, the projection estimates a subsidence rate in central Spanish

Valley of ~450 m/my (0.45 m/ky) over the last 100 ky, which may range from 380-560 m/my (0.38-0.56 m/ky) based on a 20% uncertainty in the estimated T3 age of ~100 ka.

53 This range is significantly slower than the estimated maximum rate of total Moab Valley

subsidence in the late Pleistocene of ~1000-1350 m/my (1.00-1.35 m/ky).

DISCUSSION

Controls and correlations of alluvial deposits

The chronostratigraphy of terrace, upland, and undifferentiated gravel deposits in

the Mill-Pack watersheds reflects the influences of regional climate variability, baselevel

fall from incision of the Colorado River system, and local salt-dissolution subsidence.

The T1 ages of ~16-14 ka in Mill and Pack creeks are at the end of Marine Isotope State

(MIS) 2 and suggest fluvial deposition during the decline of the Rocky Mountain

Pinedale glaciation. The height of this glacial episode has been constrained at ~19.4 ka in

the San Juan Mountains of Colorado (Guido et al., 2007), 21.4-16.5 ka in the central

Colorado Rockies (Brugger, 2007), 18-16.5ka in the southern Uinta Mountains (Laabs et al., 2009), and 23.1-20.0 ka on Boulder Mountain in the west-central Colorado Plateau

(Marchetti et al., 2005). The Pinedale-age T1 in Mill and Pack creeks correlates to the

tributary S2 terrace in eastern Grand Canyon (Anders et al., 2005; DeJong, 2007), but does not correlate with nearby terrace records on the mainstem Colorado River (Jochems

and Pederson, 2015). The T1y age of ~6 ka in Spanish Valley represents a Holocene episode of deposition on the Pack Creek fan during an otherwise incisional regime.

The three T2 OSL ages span a relatively wide range of ~78-57 ka and do not closely correspond between Mill and Pack creeks. At 57.06 ± 6.81 ka, the Mill Creek T2 is much younger than the Pack Creek T2 ages of 66.50 ± 7.45 ka and 78.22 ± 10.36 ka, which themselves also span a wide range. This wide spectrum of T2 ages within Pack

Figure 2.16. Pack Creek longitudinal profile and terrace deformation diagram. Profile spans from Brumley Creek to the deepest alluvium in Spanish Valley (Lowe et al., 2007). Numerical dates and rates reported with 1σ uncertainty and are discussed in the text. Pink dots show survey locations used to measure tilting of T3 tread. Geology shown conceptually: Km = Mancos Shale, Kd/Kbc = Dakota Sandstone and Burro Canyon Formation, Jm = Morrison Formation, Jmid = middle Jurassic section, Jn = Navajo Sandstone, Jk = Kayenta Formation.

55 Creek and between Pack and Mill creeks indicates that deposition near the edge of the

graben can be spatially and temporally heterogenous and is likely influenced by local and

potentially independent subsidence episodes in Moab and Spanish valleys. Even so, the

large T2 age range may still record an overarching climate influence. The ~78-57 ka ages

span the interglacial-glacial transition between MIS 5a and MIS 4 and extend through the

first half of MIS 3, which was a period of abrupt, millennial-scale climate variability in the southwestern U.S. (Wagner et al., 2010). These periods of highly variable climate that are neither full-glacial nor full-interglacial have been correlated with pulses of sediment production and mobilization over the late Pleistocene in the Canyonlands region, and may relate to the effects of vegetation destabilization (Reheis et al., 2005).

Similarly, the two T3 depositional ages in Mill Creek cluster tightly between 82 and 81 ka and correspond to MIS 5a, which was an intermediate phases in the global climate record that occurred after the last full interglacial of MIS 5e. Thus, similar to the

T2, this period of deposition may represent increased tributary sediment production in response to highly variable climate conditions. Though it lacks numeric age control, the

Pack Creek T3 is likely older than the ~81 ka age from Mill Creek based on the fact that the Pack Creek T2 is at least 9 ky older than the Mill Creek T2. Given the prevalence of

~100-140 ka ages for undifferentiated alluvium in the graben floor and the documented downward projection of T3 terrace treads along Pack Creek, we estimate that the Pack

Creek T3 dates to ~100 ka and is roughly age-equivalent with these Qagu deposits.

Terrace ages along Mill and Pack creeks generally do not correlate with the terrace record from regional mainstem rivers. Jochems and Pederson (2015) document three terrace levels (M2-M4) that date to between ~25 ka and ~58 ka 30 km upstream of

56 Moab Valley on the Colorado River, while the Mill Creek T2 (USU-2498) is the only sample that falls within this age range in our study area. Similarly, the piedmont-5 (~85 ka, synchronous with the Mill Creek T2) is the only deposit along the Green River at

Crystal Geyser that correlates in age to any of the terraces along Mill or Pack creeks

(Pederson et al., 2013b). This pattern corroborates the observations of Anders et al.

(2005) in eastern Grand Canyon, who found tributary aggradation to be asynchronous

with mainstem terrace formation. These results thus support the interpretation that semi-

arid tributaries have independent responses to climate change and disturbance in their

own catchments, distinct in time from the glacial outwash periods that influence terrace

formation on arterial watercourses like the Colorado River.

The five Qagu luminescence ages, sampled from one natural outcrop and one

gravel pit, range from ~137 to ~106 ka. This relatively narrow age window spans ~19 m

of stratigraphic thickness below grade in Spanish Valley and a thickness of ~16 m above

grade in Moab Valley. This suggests a major depositional episode that occurred

immediately prior to T4 time in Mill Creek, and which may have overlapped with T3

time in Pack Creek. That this period is not in-sequence with the deposition of upstream

strath terraces in Mill Creek suggests against an upstream sediment production control.

Instead, we interpret that undifferentiated alluvial gravels of this age are preferentially

preserved due to increased accommodation space, recording a period of accelerated

subsidence in Moab and Spanish valleys. If the Pack Creek T3 does in fact have an age of

~100 ka that correlates with the younger Qagu ages, it may reflect the fact that fill

terraces in southeast Spanish Valley are more influenced by subsidence and

57 accommodation space than are the strath terraces immediately outside the graben margin

along Mill Creek.

An earlier episode of extensive aggradation and relatively steady baselevel is

recorded in the ~1.7-1.6 Ma T7 upland gravels of South Mountain and Johnson’s Ridge.

These 20-50 m thick gravel deposits are interpreted as remnants of an extensive piedmont bajada that today covers the north, east, and south flanks of the La Sal Mountains

(Doelling, 2001, 2004). An alternative explanation is that gravels on Johnson’s Ridge are shoestring deposits from a paleo-South Mill Creek canyon that represent inverted topography, though the balance of field evidence suggests that the connected-piedmont hypothesis is a simpler explanation (Chapter 3). The large spatial extent and thickness of these gravels suggests that they represent an early-Pleistocene period of baselevel stability and net sediment storage. This timespan appears to post-date the Miocene to middle-Pliocene period of slow regional exhumation recognized from thermochronlogy

(Murray et al., 2016; Murray et al., in review), and its cause remains uncertain. In summary, different temporal controls are interpreted for the deposition of strath terraces along Mill Creek, fill terraces in Pack Creek together with undifferentiated gravels in

Moab and Spanish valleys, and the Johnson’s Ridge/South Mountain upland gravels. The depositional ages of these gravels appear to be most-closely coupled to climate and sediment production in the La Sal Mountain headwaters, subsidence-driven accommodation space in the Moab-Spanish Valley graben, and regional baselevel patterns, respectively.

58 Graben subsidence and faulting informed by deformed fluvial terraces

Surveying and geochronological correlations demonstrate that terraces on the

margins of Moab and Spanish valleys are indeed deformed (Figs. 2.14 and 2.16), as

hypothesized by previous workers (Richmond, 1962; Harden et al., 1985; Guerrero et al.,

2015). This Quaternary deformation is distinct between Mill Creek, where terraces are

abruptly truncated across the Kayenta Heights fault zone, and Pack Creek, where

upstream terraces are down-warped into the basin floor over the course of 5 kilometers.

Rather than representing fundamentally different processes of subsidence between Moab

and Spanish valleys, these deformation styles reflect the different orientations of the two

drainages as they enter the graben, with Mill Creek crossing the margin perpendicular to

strike and Pack Creek entering the graben along its axis.

In Moab Valley, chronostratigraphy at Mill Creek constrains slip across the

Kayenta Heights fault zone to 439 ± 49 m/my (0.439 ± 0.049 m/ky) over the last ~200 ky. Based on the relative heights above grade of terraces at the Mill Creek trailhead and alluvium beneath the Colorado River, the maximum total subsidence in Moab Valley is estimated at ~1000-1350 m/my (1.00-1.35 m/ky) for the middle to late Pleistocene. This range of subsidence rates, which assumes a T5-age for the deepest alluvium in Moab

Valley, is preferred over much slower and longer time-period rates that assume the basal alluvium is T7-age. This is because the preservation potential of early-Pleistocene gravels in the bottom of Moab Valley is much lower than for younger T5-age deposits, and it is

possible that any early-Pleistocene gravels were eroded from the valley floor prior to the

~100 ka period of aggradation recorded by the gravel pit and bowling alley samples. If

59 the estimates of alluvium depth from boreholes are correct, and if the deepest gravels in

Moab Valley are indeed younger than ~200 ka, then this net subsidence rate is at least

double the slip on the KHFZ, suggesting that much of Moab Valley’s subsidence is

accommodated basin-ward of the mapped fault zone. This excess subsidence may occur across additional unmapped salt-dissolution faults similar in style to the KHFZ, or through gradually intensifying dissolution of the Paradox Formation caprock towards the

Colorado River – similar to patterns along Pack Creek in Spanish Valley. Though speculative, the lack of documented fault scarps in the alluvium of Moab Valley makes the passive down-warping mechanism the simpler explanation.

The middle-to-late-Pleistocene Kayenta Heights fault zone slip rate varies

spatially and temporally with other independent estimates of displacement. It is an order

of magnitude slower than the short term late-Holocene slip rate of >3000 m/my (>3 m/ky) that Guerrero et al. (2015) calculated for a strand of the KHFZ just 3.5 km southeast of Mill Creek. This slip rate is based on geochronology from a trench spanning the KHFZ with absolute ages between 4.5 and 2.3 ka, and likely captures a period of rapid local deformation set within the spatially and temporally heterogeneous fault zone

(Chapter 4, Plate 3). On the other side of the spectrum, it is apparent that the longer-term average subsidence since the early Pleistocene in Moab Valley is slower than the middle- to-late-Pleistocene displacement rate across the KHFZ. A GIS reconstruction and projection of the T7 paleo-piedmont inferred to exist before the initiation of graben subsidence suggests that a maximum of ~280 m of subsidence has occurred in Moab

Valley since the early Pleistocene (Chapter 3). Given the 439 ± 49 m/my (0.439 ± 0.049 m/ky) slip rate along the KHFZ, this amount of subsidence could have been

60 accomplished in just ~640 ky if displacement along the KHFZ occurred at a uniform rate.

Yet cosmogenic burial dating of the Johnson’s Ridge and South Mountain T7 gravels

suggests that these surfaces were abandoned much earlier and that the calculated late-

Pleistocene subsidence and faulting rates cannot have been sustained through the entire

Pleistocene. Taken together, this suggests that subsidence in Moab Valley was modest in

the early Pleistocene before accelerating sometime closer to the present.

In contrast to Mill Creek, Pack Creek terraces appear to have experienced gradual down-warping through the Quaternary rather than deformation across faults. The subsidence recorded by the downward projection of terrace treads is ~450 m/my (0.45 m/ky) over the last ~100 ky in the center of Spanish Valley. Similar to the Moab Valley case, the maximum depth of subsidence in Spanish Valley since T7 time informed by the reconstructed paleo-piedmont is 265 m (Chapter 3). At the calculated late-Pleistocene

subsidence rate this depth of lowering would have occurred in only ~590 ky, which is

much less time than the inferred ~1.5 my since T7 abandonment. Thus, as in Moab

Valley, there is evidence in Spanish Valley that subsidence in the early Pleistocene must

have been slower that it has been over the late Pleistocene and Holocene.

Relief production in the Moab-Spanish Valley salt graben

Geochronology and mapping (Chapter 4) suggest that the topographic relief along

the northeast graben margin may have been produced entirely in the Quaternary. The two

~1.7-1.6 Ma cosmogenic burial dates on long-lived upland gravels have important

implications for the initiation of graben subsidence. The preserved upland gravel at

Johnson’s Ridge in particular represents the past position of a drainage that transported

61 sediment from the La Sal Mountains. It is reasonable to assume that this paleo-drainage would seek the lowest position in the landscape over which to flow, so there could not have been a large valley to the southwest at the time that the drainage deposited the

Johnson’s Ridge gravel. This suggests that the 240 m of basin-margin relief in modern times between Johnson’s Ridge and the bottom of Spanish Valley near Ken’s Lake must have been produced after ~1.5 Ma, the estimated age for the upland gravel tread (Chapter

3). Because there are no higher remnant gravels or escarpments on the northeast side of

Spanish Valley, there is no evidence for faulting and graben subsidence pre-dating the

South Mountain/Johnson’s Ridge gravels.

The reconstruction and projection of this correlative T7 surface in Chapter 3 indicates that there has been a spatially-averaged ~130 m of surface lowering below it in the Moab-Spanish Valley graben. This topographic reconstruction also shows the likelihood of an escarpment during T7 time along the southwest margin of Moab and

Spanish valleys analogous to the modern-day Moab Rim (Fig. 2.1). It is probable that the escarpment was farther northeast during T7 time, and has since retreated to the southwest with the initiation of graben subsidence. This escarpment would have had less topographic relief than today, with the toe of the projected T7 paleo-piedmont extending to its base (Chapter 3, Fig. 3.8). Regardless, rather than indicating an early expression of the salt graben, this escarpment is taken to represent paleo-relief on the Moab fault similar to that seen today northwest of Moab Valley (Chapter 4; Foxford et al., 1998).

62 Controls on the subsidence of Moab and Spanish valleys

Moab and Spanish valleys appear to have evolved as separate subsidence centers within the larger salt graben. The Pack Creek T2 terrace reappears across the topographic saddle between the two valleys, and the presence of exposed bedrock indicates this region of the graben has undergone less subsidence than the neighboring alluvial valleys. The valley-bounding topographic saddle coincides with a segment boundary along the northeast graben margin in the Kayenta Heights fault zone (Chapter 4). Furthermore, the bedrock architecture beneath the two basins is dissimilar, with Moab Valley underlain directly by Paradox Formation caprock and Spanish Valley underlain by Jurassic strata

(Doelling et al., 2002; Lowe et al., 2007). It is uncertain when post-Paradox bedrock was eroded from the top of the anticline over what is now Moab Valley (Doelling et al.,

2002), but it likely occurred when this strata occupied a higher position in the landscape before the initiation of anticline collapse. Finally, chronostratigraphy and estimated alluvium thicknesses suggest that subsidence in Moab Valley has potentially occurred twice as rapidly as in Spanish Valley over the last ~200 ky, which is consistent with these two basins subsiding separately.

Though Moab and Spanish valleys represent distinct depocenters that may be subsiding at different rates, they share the same processes of salt dissolution and a common regional baselevel. Subsidence in Moab Valley in the last ~200 ky (Fig. 2.14) has been faster than its average over the last ~1.5 my (Chapter 3). A similar rate acceleration prior to ~200 ka is recorded in upstream tributary incision based on the chronostratigraphy of South Mill Creek at RK 19 (Fig. 2.13A). That subsidence in Moab

63 Valley may have accelerated at the same time as incision along Mill Creek suggests a

common control, which is likely regional baselevel fall on the Colorado River (Chapter

3). Though they may experience parallel rate changes through time, the processes of

graben subsidence and regional incision do not necessarily occur at the same absolute

rates. This is illustrated by the maximum subsidence rate in Moab Valley of ~1000-1350

m/my (1.00-1.35 m/ky) that well exceeds regional Colorado River incision rates of 300-

900 m/my (0.3-0.9 m/ky) (Jochems and Pederson, 2015; Pederson, personal

communication). In fact, subsidence in the last ~200 ky must have been more rapid than

local mainstem and tributary incision rates for over 125 m of middle-to-late-Pleistocene alluvium to be deposited beneath northwest Moab Valley (Doelling et al., 2002).

These subsidence and incision relations are possible because groundwater dissolution is hypothesized to drive the active subsidence and deformation of this region

(Huntoon, 1982; Doelling et al., 1988; Gutiérrez, 2004). As the Colorado River incises, the regional water table lowers and salt dissolution is expected to penetrate greater stratigraphic depths. Thus, it is not the Colorado River itself but the indirectly-connected depth of groundwater infiltration that drives salt-dissolution subsidence in Moab and

Spanish Valleys, allowing subsidence rates to track with, but generally exceed, river incision. Given this mechanism, it is logical that Moab Valley is subsiding more rapidly than Spanish Valley on account of its low elevation at the Colorado River, the downward hydrologic gradient between Spanish Valley and Moab Valley, and the groundwater discharge into springs and wetlands in lower Moab Valley (Sumsion, 1971; Blanchard,

1990; Lowe et al., 2007). Furthermore, these processes imply that regional baselevel signals from the Colorado River can be transmitted across the graben to bedrock reaches

64 of Mill and Pack creeks intact and unamplified. This is because any excess

accommodation space from subsidence is erased by depositional filling as these drainages

maintain grade with the outlet of the Colorado River from Moab Valley. Thus, it is

interpreted that regional baselevel fall on the Colorado River controls both graben

subsidence and tributary incision: subsidence indirectly through groundwater infiltration

and dissolution, and incision through the direct propagation of regional baselevel signals

across the lower graded reaches of Mill and Pack creeks.

Incision rates and implications for regional landscape evolution

The results offer compelling evidence of a wave of transient incision propagating through Mill and Pack creeks. In upper South Mill Creek the elevation and ages of terrace and upland gravels reveals a two-stage incision history, where stability and modest incision between the T7 and T5 levels occurred at 0-50 m/my (0-0.05 m/ky) over the early Pleistocene before incision accelerated to at least 619 ± 80 m/my (0.619 ± 0.080 m/ky) between T6 and T5 time (Fig. 2.13A). Late-Pleistocene incision below the T5 has occurred at a similar rate in the lower Mill Creek trailhead reach (Fig. 2.13B), which also falls within the range of nearby incision rates calculated on the Colorado River (Jochems and Pederson, 2015; Pederson, personal communication). Though the Pack Creek terrace record is less well preserved and dated, a similar two-stage history is evident with incision since ~100 ka calculated at 501 ± 65 m/my (0.501 ± 0.065 m/ky) while the older

T7 upland gravels reflect a period of baselevel stability (Fig. 2.15). Transient incision in

Mill Creek is also supported by strath terraces that form at knickpoints and grow higher downstream, which is documented in the T1, T2, and T3 (Fig. 2.14) (Zaprowski et al.,

65 2001; Pazzaglia, 2013). In the ideal case, ages from a single terrace level would grow

younger upstream to reflect more recent abandonment as the knickpoint migrates

headward (Zaprowski et al., 2001; Pazzaglia, 2013), though this signal is likely obscured

by ~10 ky error bars for the T3 ages in Mill Creek. The baselevel fall signal responsible

for this transient incision must have arrived in the Mill-Pack creek watershed between T6

and T5 time, as T6 gravels form broad upland surfaces whereas T5 deposits are strath

terraces reflecting confinement in a bedrock canyon. There is no age control on the T6,

but it is inferred that the wave of transient incision arrived in the system closer to T5 time

(inspiring a prior-to-200 ka estimate) on account of the long-term stability implied by the

Johnsons Up On Top upland gravels. Since the abandonment of the T6 upland gravels

and the beginning of more rapid baselevel fall, this transient wave has resulted in ~150 m

of incision in the canyons of Mill Creek.

The source of baselevel fall driving transient incision was likely the Colorado

River rather than graben subsidence considering that subsidence itself is indirectly controlled by fluvial incision and groundwater infiltration. Graben subsidence is interpreted to function only as a local baselevel for Mill and Pack creeks, with its signal muted by the depositional filling of the basin as tributaries remain graded to the Colorado

River. Baselevel fall from the Colorado River is recorded by major knickzones in resistant lithology that occupy similar positions in drainage networks throughout the

Moab region, including outside of the Moab-Spanish Valley graben (Chapter 3). In the forks of Mill and Pack creeks, these major knickzones are between 100 and 200 m tall

(Chapter 3, Fig. 3.2), suggesting that they may represent the head of this transient wave of incision, currently stalled on resistant bedrock, that has been working up the drainage

66 network over the late Pleistocene and has left ~150 m deep canyons in its wake. In total,

this evidence inspires a conceptual model for the central Colorado Plateau where signals

of baselevel fall on the Colorado River are transmitted upstream across salt structures,

which deform at rates indirectly governed by these same baselevel signals and thus act as

conduits rather than barriers to the migration of transient incision.

The episode of Colorado River baselevel fall responsible for the transient incision

documented in Mill and Pack creeks remains unconstrained. The ~5 Ma integration of the

Colorado River though the Grand Canyon constituted a major baselevel fall, and initiated

transient incision that has since been working upstream (e.g. Pederson et al., 2002). It is unknown whether this transient incision signal has propagated to the central Colorado

Plateau. Pelletier (2010) estimated from modeling that the knickpoint at the head of this transient wave migrated through the Grand Canyon at a rate of ~100 km/my, and

Bursztyn et al. (2015) suggest that the Lee’s Ferry knickpoint is pinned by bedrock such that any transience must be farther upstream. Others have proposed that transient incision

has reached tributaries upstream of Grand Canyon in the last ~500 ky (Cook et al., 2009;

Darling et al., 2012). Whether the prior-to-200 ka arrival of transient incision recorded by terrace and upland gravels in Mill and Pack creeks is related to the integration of the

Colorado River system or a more recent baselevel fall remains unconstrained. Regardless, this work adds to the growing body of evidence of youthful, transient incision in the central Colorado Plateau.

67 Summary

Chronostratigraphy of alluvial deposits in the watersheds of the Moab-Spanish

Valley salt graben reveals seven levels of terraces and upland gravels, as well as a valley-

bottom alluvium. Temporal patterns of deposition reflect the independent influences of

local climate variability and sediment production, regional baselevel stability, and

changes in accommodation space driven by basin subsidence. Terraces along Mill and

Pack creeks are deformed at the margin of the salt graben, and their age and surveyed

elevations document Quaternary subsidence. Mill Creek terraces are offset at the graben

margin and record 439 ± 49 m/my (0.439 ± 0.049 m/ky) of displacement across the

Kayenta Heights fault zone in the middle to late Pleistocene, while maximum total

subsidence in Moab Valley over this same period may be as rapid as ~1000-1350 m/my

(1.00-1.35 m/ky). As Pack Creek enters the graben axially in southeast Spanish Valley its

terraces are down-warped by 0.9° into the basin floor, the projection of which suggests

~450 m/my (0.45 m/ky) of subsidence in the last 100 ky. Spatial relations of a

reconstructed T7-age piedmont (Chapter 3) suggest that the measured post-200 ka

subsidence rates could not have been sustained through the entire Pleistocene and that

subsidence instead accelerated sometime in the middle to late Pleistocene. Furthermore,

the preservation of remnant upland gravels along the rim of Spanish Valley imply that the

modern relief of the northeast graben margin has been produced since ~1.5 Ma and that

graben subsidence is a Quaternary phenomenon.

Terrace and upland gravels in Mill and Pack creeks record compelling evidence

for transient incision, exhibiting a two-stage history with little to no incision through the

early Pleistocene followed by higher incision rates of ~530-620 m/my (0.53-0.62 m/ky)

68 since ~200 ka. The documented acceleration of subsidence in Moab and Spanish valleys may track with the prior-to-200 ka acceleration of incision, suggesting that salt dissolution is coupled to baselevel fall along the Colorado River through groundwater infiltration. This wave of incision is interpreted to be currently stalled on resistant lithology in the upper reaches of the Mill-Pack creek watershed. The transient incision recorded in these drainages may relate to the ~5 Ma integration of the Colorado River system or may reflect a more recent and poorly-understood pulse of baselevel fall. This work demonstrates that despite the complicating influence of local and active salt deformation, baselevel signals from far downstream have worked their way into tributary drainages in the central Colorado Plateau and are recorded in the spatial and temporal distribution of alluvial gravels.

REFERENCES

Anders, M.D., Pederson, J.L., Rittenour, T.M., Sharp, W.D., Gosse, J.C., Karlstrom, K.E., Crossey, L.J., Goble, R.J., Stockli, L., and Yang, G., 2005, Pleistocene geomorphology and geochronology of eastern Grand Canyon: linkages of landscape components during climate changes: Quaternary Science Reviews, v. 24, p. 2428–2448, doi: 10.1016/j.quascirev.2005.03.015.

Aslan, A., and Hanson, P., 2009, Late Pleistocene Colorado River terraces, western Colorado: A test of the stream power model (abstract), in Geological Society of America Abstracts with Programs, v. 41(7), p. 622.

Baars, D.L., and Doelling, H.H., 1987, Moab salt-intruded anticline, east-central Utah: Geological Society of America Centennial Field Guide, Rocky Mountain Section, p. 275–280.

Balco, G., and Rovey, C.W., 2008, An isochron method for cosmogenic-nuclide dating of buried soils and sediments: American Journal of Science, v. 308, p. 1083–1114.

Barbeau, D.L., 2003, A flexural model for the Paradox Basin: implications for the tectonics of the Ancestral Rocky Mountains: Basin Research, v. 15, p. 97–115, doi: 10.1046/j.1365-2117.2003.00194.x.

69 Best, J.L., and Ashworth, P.J., 1997, Scour in large braided rivers and the recognition of sequence stratigraphic boundaries: Nature, v. 387, p. 275.

Biggar, N.E., and Adams, J.A., 1987, Dates Derived from Quaternary Strata in the Vicinity of Canyonlands National Park, in Field Symposium, Guidebook of the Four Corners Geological Society, v. 10, p. 127–136, http://archives.datapages.com/data/fcgs/data/018/018001/127_four- corners180127.htm (accessed December 2017).

Black, B.D., Hylland, M.D., and Hecker, S., 2004, Fault number 2476, Moab fault and deformation zones: Quaternary fault and fold database of the United States, https://earthquake.usgs.gov/cfusion/qfault/show_report_AB_archive.cfm?fault_id =2476§ion_id= (accessed February 2018).

Blanchard, P.J., 1990, Ground-water conditions in the Grand County area, Utah, with emphasis on the Mill Creek-Spanish Valley area: Utah Department of Natural Resources, Division of Water Rights Technical Publication 100, 69 p., https://pubs.er.usgs.gov/publication/70179034 (accessed January 2018).

Brugger, K.A., 2007, Cosmogenic 10Be and 36Cl ages from Late Pleistocene terminal moraine complexes in the Taylor River drainage basin, central Colorado, USA: Quaternary Science Reviews, v. 26, p. 494–499, doi: 10.1016/j.quascirev.2006.09.006.

Bull, W.B., 1991, Geomorphic responses to climatic change: New York, Oxford University Press, https://www.osti.gov/scitech/biblio/5603696 (accessed December 2017).

Bursztyn, N., Pederson, J.L., Tressler, C., Mackley, R.D., and Mitchell, K.J., 2015, Rock strength along a fluvial transect of the Colorado Plateau – quantifying a fundamental control on geomorphology: Earth and Planetary Science Letters, v. 429, p. 90–100, doi: 10.1016/j.epsl.2015.07.042.

Caldwell, J., and Vahidsafa, A., 2013, Propagation of Error: Chemistry LibreTexts, https://chem.libretexts.org/Core/Analytical_Chemistry/Quantifying_Nature/Signif icant_Digits/Propagation_of_Error (accessed April 2018).

Carbonel, D., Gutiérrez, F., Linares, R., Roqué, C., Zarroca, M., McCalpin, J., Guerrero, J., and Rodríguez, V., 2013, Differentiating between gravitational and tectonic faults by means of geomorphological mapping, trenching and geophysical surveys. The case of the Zenzano Fault (Iberian Chain, N Spain): Geomorphology, v. 189, p. 93–108, doi: 10.1016/j.geomorph.2013.01.020.

Cater, F.W., 1970, Geology of the salt anticline region in southwestern Colorado: U.S. Geological Survey Professional Paper 637, 80 p., https://pubs.usgs.gov/pp/0637/report.pdf.

70 Colman, S.M., 1983, Influence of the Onion Creek salt diapir on the late Cenozoic history of Fisher Valley, southeastern Utah: Geology, v. 11, p. 240–243, doi: 10.1130/0091-7613(1983)11<240:IOTOCS>2.0.CO;2.

Cook, K.L., Whipple, K.X., Heimsath, A.M., and Hanks, T.C., 2009, Rapid incision of the Colorado River in Glen Canyon – insights from channel profiles, local incision rates, and modeling of lithologic controls: Earth Surface Processes and Landforms, v. 34, p. 994–1010, doi: 10.1002/esp.1790.

Darling, A.L., Karlstrom, K.E., Granger, D.E., Aslan, A., Kirby, E., Ouimet, W.B., Lazear, G.D., Coblentz, D.D., and Cole, R.D., 2012, New incision rates along the Colorado River system based on cosmogenic burial dating of terraces: Implications for regional controls on Quaternary incision: Geosphere, v. 8, p. 1020–1041, doi: 10.1130/GES00724.1.

DeJong, B.D., 2007, Effects of Quaternary climate change on tributary sedimentation and geomorphology in eastern Grand Canyon [M.S. thesis]: Utah State University, 122 p.

Ditchburn, R.G., and Whitehead, N.E., 1994, The separation of 10Be from silicates, in 3d Workshop of the South Pacific Environmental Radioactivity Association, p. 4–7.

Doelling, H.H., 2004, Geologic Map of the La Sal 30’ x 60’ Quadrangle, San Juan, Wayne, and Garfield Counties, Utah, and Montrose and San Miguel Counties, Colorado: Utah Geological Survey.

Doelling, H.H., 2001, Geologic map of the Moab and eastern part of the San Rafael Desert 30’ x 60’ quadrangles, Grand and Emery Counties, Utah, and Mesa County, Colorado: Utah Geological Survey.

Doelling, H.H., Oviatt, C.G., and Huntoon, P.W., 1988, Salt deformation in the Paradox region: Utah Geological Survey, v. Bulletin 122.

Doelling, H.H., Ross, M.L., and Mulvey, W.E., 2002, Geologic Map of the Moab 7.5’ Quadrangle, Grand County, Utah: Utah Geological Survey.

Dorsey, R.J., Fluette, A., McDougall, K., Housen, B.A., Janecke, S.U., Axen, G.J., and Shirvell, C.R., 2007, Chronology of Miocene–Pliocene deposits at Split Mountain Gorge, Southern California: A record of regional tectonics and Colorado River evolution: Geology, v. 35, p. 57–60, doi: 10.1130/G23139A.1.

Faulds, J.E., Wallace, M.A., Gonzalez, L.A., Heizler, M.T., Young, R.A., and Spamer, E.E., 2001, Depositional environment and paleogeographic implications of the late Miocene Hualapai Limestone, northwestern Arizona and southern Nevada: The Colorado River: Origin and Evolution: Grand Canyon, Arizona, Grand Canyon Association Monograph, v. 12, p. 81–88.

71 Foxford, K.A., Walsh, J.J., Watterson, J., Garden, I.R., Guscott, S.C., and Burley, S.D., 1998, Structure and content of the Moab Fault Zone, Utah, USA, and its implications for fault seal prediction: Geological Society, London, Special Publications, v. 147, p. 87–103, doi: 10.1144/GSL.SP.1998.147.01.06.

Furuya, M., Mueller, K., and Wahr, J., 2007, Active salt tectonics in the Needles District, Canyonlands (Utah) as detected by interferometric synthetic aperture radar and point target analysis: 1992–2002: Journal of Geophysical Research: Solid Earth, v. 112, p. B06418, doi: 10.1029/2006JB004302.

Galbraith, R.F., and Roberts, R.G., 2012, Statistical aspects of equivalent dose and error calculation and display in OSL dating: An overview and some recommendations: Quaternary Geochronology, v. 11, p. 1–27, doi: 10.1016/j.quageo.2012.04.020.

Gallen, S.F., Pazzaglia, F.J., Wegmann, K.W., Pederson, J.L., and Gardner, T.W., 2015, The dynamic reference frame of rivers and apparent transience in incision rates: Geology, p. G36692.1, doi: 10.1130/G36692.1.

Geiger, F., 2014, Landscape evolution of the Needles Fault Zone, Utah, investigated through chronostratigraphic and terrain analysis [M.S. thesis]: Utah State University, 172 p., http://digitalcommons.usu.edu/etd/3885.

Granger, D.E., 2006, A review of burial dating methods using 26Al and 10Be: Geological Society of America Special Papers, v. 415, p. 1–16.

Granger, D.E., 2013, Cosmogenic Nuclide Burial Dating in Archaeology and Paleoanthropology: Treatise on Geochemistry: Second Edition, v. 14, p. 81–97, doi: 10.1016/B978-0-08-095975-7.01208-0.

Guérin, G., Mercier, N., and Adamiec, G., 2011, Dose-rate conversion factors: update: Ancient TL, v. 29, p. 5–8.

Guerrero, J., Bruhn, R.L., McCalpin, J.P., Gutiérrez, F., Willis, G., and Mozafari, M., 2015, Salt-dissolution faults versus tectonic faults from the case study of salt collapse in Spanish Valley, SE Utah (USA): Lithosphere, v. 7, p. 46–58, doi: 10.1130/L385.1.

Guido, Z.S., Ward, D.J., and Anderson, R.S., 2007, Pacing the post–Last Glacial Maximum demise of the Animas Valley glacier and the San Juan Mountain ice cap, Colorado: Geology, v. 35, p. 739–742, doi: 10.1130/G23596A.1.

Gutiérrez, F., 2004, Origin of the salt valleys in the Canyonlands section of the Colorado Plateau: Evaporite-dissolution collapse versus tectonic subsidence: Geomorphology, v. 57, p. 423–435, doi: 10.1016/S0169-555X(03)00186-7.

72 Gutiérrez, F., Carbonel, D., Kirkham, R.M., Guerrero, J., Lucha, P., and Matthews, V., 2014, Can flexural-slip faults related to evaporite dissolution generate hazardous earthquakes? The case of the Grand Hogback monocline of west-central Colorado: GSA Bulletin, v. 126, p. 1481–1494, doi: 10.1130/B31054.1.

Hackman, R.J., 1956, Photogeologic map of the Mount Peale-4 Quadrangle, San Juan County, Utah: U.S. Geological Survey.

Hancock, G.S., and Anderson, R.S., 2002, Numerical modeling of fluvial strath-terrace formation in response to oscillating climate: GSA Bulletin, v. 114, p. 1131–1142, doi: 10.1130/0016-7606(2002)114<1131:NMOFST>2.0.CO;2.

Harden, D.R., Biggar, N.E., and Gillam, M.L., 1985, Quaternary deposits and soils in and around Spanish Valley, Utah: GSA Special Papers, v. 203, p. 43–64, doi: https://dx.doi.org/10.1130/SPE203.

Howard, K.A., House, P.K., Dorsey, R.J., and Pearthree, P.A., 2015, River-evolution and tectonic implications of a major Pliocene aggradation on the lower Colorado River: The Bullhead Alluvium: Geosphere, v. 11, p. 1–30, doi: 10.1130/GES01059.1.

Humphreys, E., Hessler, E., Dueker, K., Farmer, G.L., Erslev, E., and Atwater, T., 2003, How Laramide-Age Hydration of North American Lithosphere by the Farallon Slab Controlled Subsequent Activity in the Western United States: International Geology Review, v. 45, p. 575–595, doi: 10.2747/0020-6814.45.7.575.

Hunt, C.B., 1956, Cenozoic geology of the Colorado Plateau: U.S. Geological Survey Professional Paper 279, 99 p., http://pubs.er.usgs.gov/publication/pp279 (accessed October 2017).

Huntley, D.J., Godfrey-Smith, D.I., and Thewalt, M.L.W., 1985, Optical dating of sediments: Nature, v. 313, p. 105, doi: 10.1038/313105a0.

Huntoon, P.W., 1999, Field-based identification of salt-related structures and their differentiation from tectonic structures: Techniques for Identifying Faults and Determining their Origins: Washington, DC, US Nuclear Regulatory Commission, contract report NUREG/CR-5503.

Huntoon, P.W., 1982, The Meander anticline, Canyonlands, Utah: An unloading structure resulting from horizontal gliding on salt: GSA Bulletin, v. 93, p. 941–950, doi: 10.1130/0016-7606(1982)93<941:TMACUA>2.0.CO;2.

Hütt, G., Jaek, I., and Tchonka, J., 1988, Optical dating: K-feldspars optical response stimulation spectra: Quaternary Science Reviews, v. 7, p. 381–385, doi: 10.1016/0277-3791(88)90033-9.

73 Jochems, A.P., and Pederson, J.L., 2015, Active salt deformation and rapid, transient incision along the Colorado River near Moab, Utah: Journal of Geophysical Research: Earth Surface, v. 120, p. 2014JF003169, doi: 10.1002/2014JF003169.

Karlstrom, K.E., Coblentz, D., Dueker, K., Ouimet, W., Kirby, E., Wijk, J.V., Schmandt, B., Kelley, S., Lazear, G., Crossey, L.J., Crow, R., Aslan, A., Darling, A., Aster, R., et al., 2012, Mantle-driven dynamic uplift of the Rocky Mountains and Colorado Plateau and its surface response: Toward a unified hypothesis: Lithosphere, v. 4, p. 3–22, doi: 10.1130/L150.1.

Laabs, B.J.C., Refsnider, K.A., Munroe, J.S., Mickelson, D.M., Applegate, P.J., Singer, B.S., and Caffee, M.W., 2009, Latest Pleistocene glacial chronology of the Uinta Mountains: support for moisture-driven asynchrony of the last deglaciation: Quaternary Science Reviews, v. 28, p. 1171–1187, doi: 10.1016/j.quascirev.2008.12.012.

Lazear, G., Karlstrom, K., Aslan, A., and Kelley, S., 2013, Denudation and flexural isostatic response of the Colorado Plateau and southern Rocky Mountains region since 10 Ma: Geosphere, v. 9, p. 792–814, doi: 10.1130/GES00836.1.

Lee, J.P., Stockli, D.F., Kelley, S.A., Pederson, J.L., Karlstrom, K.E., and Ehlers, T.A., 2013, New thermochronometric constraints on the Tertiary landscape evolution of the central and eastern Grand Canyon, Arizona: Geosphere, v. 9, p. 216–228, doi: 10.1130/GES00842.1.

Levander, A., Schmandt, B., Miller, M.S., Liu, K., Karlstrom, K.E., Crow, R.S., Lee, C.- T.A., and Humphreys, E.D., 2011, Continuing Colorado plateau uplift by delamination-style convective lithospheric downwelling: Nature, v. 472, p. 461– 465, doi: 10.1038/nature10001.

Lowe, M., Wallace, J., Kirby, S.M., and Bishop, C.E., 2007, The hydrogeology of Moab- Spanish Valley, Grand and San Juan Counties, Utah, with emphasis on maps for water-resource management and land-use planning: Utah Geological Survey Special Study 120, http://digitallibrary.utah.gov/awweb/awarchive?item=31274.

Marchetti, D.W., Cerling, T.E., and Lips, E.W., 2005, A glacial chronology for the Fish Creek drainage of Boulder Mountain, Utah, USA: Quaternary Research, v. 64, p. 264–271, doi: 10.1016/j.yqres.2005.05.004.

Matmon, A., Stock, G.M., Granger, D.E., and Howard, K.A., 2012, Dating of Pliocene Colorado River sediments: Implications for cosmogenic burial dating and the evolution of the lower Colorado River: GSA Bulletin, v. 124, p. 626–640, doi: 10.1130/B30453.1.

McCalpin, J.P., 2009, Chapter 2A Field Techniques in Paleoseismology—Terrestrial Environments, in International Geophysics, Academic Press, Paleoseismology, v. 95, p. 29–118, doi: 10.1016/S0074-6142(09)95002-1.

74 McKnight, E.T., 1940, Geology of area between Green and Colorado rivers, Grand and San Juan counties, Utah: U.S. Geological Survey Bulletin 908, 147 p., https://pubs.er.usgs.gov/publication/b908.

Merritts, D.J., Vincent, K.R., and Wohl, E.E., 1994, Long river profiles, tectonism, and eustasy: A guide to interpreting fluvial terraces: Journal of Geophysical Research: Solid Earth, v. 99, p. 14031–14050, doi: 10.1029/94JB00857.

Morgan, P., and Swanberg, C.A., 1985, On the Cenozoic uplift and tectonic stability of the Colorado Plateau: Journal of Geodynamics, v. 3, p. 39–63, doi: 10.1016/0264- 3707(85)90021-3.

Moucha, R., Forte, A.M., Rowley, D.B., Mitrovica, J.X., Simmons, N.A., and Grand, S.P., 2009, Deep mantle forces and the uplift of the Colorado Plateau: Geophysical Research Letters, v. 36, doi: http://dx.doi.org/10.1029/2009GL039778.

Murray, A.S., and Olley, J.M., 2002, Precision and accuracy in the optically stimulated luminescence dating of sedimentary quartz: a status review: Geochronometria, v. 21, p. 1–16.

Murray, K.E., Reiners, P.W., and Thomson, S.N., 2016, Rapid Pliocene–Pleistocene erosion of the central Colorado Plateau documented by apatite thermochronology from the Henry Mountains: Geology, v. 44, p. 483–486, doi: 10.1130/G37733.1.

Murray, K.E., Reiners, P.W., Thomson, S.N., Robert, X., and Whipple, K.X., in review, The thermochronologic record of erosion and magmatism in the Canyonlands region of the Colorado Plateau: American Journal of Science.

Murray, A.S., and Wintle, A.G., 2000, Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol: Radiation Measurements, v. 32, p. 57– 73, doi: 10.1016/S1350-4487(99)00253-X.

Nelson, S.T., Heizler, M.T., and Davidson, J.P., 1992, New 40Ar/39Ar Ages of Intrusive Rocks from the Henry and La Sal Mountains, Utah: Utah Geological Survey, 98 p.

Nuccio, V.F., and Condon, S.M., 1996, Burial and thermal history of the Paradox Basin, Utah and Colorado, and petroleum potential of the middle Pennsylvanian Paradox Formation: U.S. Geological Survey Bulletin 2000-O, 41 p., https://pubs.usgs.gov/bul/b2000o/b2000o.pdf (accessed December 2017).

Pazzaglia, F.J., 2013, 9.22 Fluvial terraces, in Shroder, J.F. ed., Treatise on Geomorphology, San Diego, CA, Academic Press, p. 379–412.

75 Pederson, J.L., Anders, M.D., Rittenhour, T.M., Sharp, W.D., Gosse, J.C., and Karlstrom, K.E., 2006, Using fill terraces to understand incision rates and evolution of the Colorado River in eastern Grand Canyon, Arizona: Journal of Geophysical Research: Earth Surface, v. 111, p. F02003, doi: 10.1029/2004JF000201.

Pederson, J.L., Burnside, N., Shipton, Z., and Rittenour, T., 2013b, Rapid river incision across an inactive fault—Implications for patterns of erosion and deformation in the central Colorado Plateau: Lithosphere, v. 5, p. 513–520, doi: 10.1130/L282.1.

Pederson, J.L., Cragun, W.S., Hidy, A.J., Rittenour, T.M., and Gosse, J.C., 2013a, Colorado River chronostratigraphy at Lee’s Ferry, Arizona, and the Colorado Plateau bull’s-eye of incision: Geology, v. 41, p. 427–430, doi: 10.1130/G34051.1.

Pederson, J.L., Mackley, R.D., and Eddleman, James L., 2002, Colorado Plateau uplift and erosion evaluated using GIS: GSA Today, v. 12, p. 4–10.

Pelletier, J.D., 2010, Numerical modeling of the late Cenozoic geomorphic evolution of Grand Canyon, Arizona: GSA Bulletin, v. 122, p. 595–608, doi: 10.1130/B26403.1.

Raup, O.B., and Hite, R.J., 1992, Lithology of evaporite cycles and cycle boundaries in the upper part of the Paradox Formation of the Hermosa Group of Pennsylvanian age in the Paradox Basin, Utah and Colorado: U.S. Geological Survey Bulletin 2000-B, 37 p., https://pubs.usgs.gov/bul/2000b/report.pdf.

Reheis, M.C., Reynolds, R.L., Goldstein, H., Roberts, H.M., Yount, J.C., Axford, Y., Cummings, L.S., and Shearin, N., 2005, Late Quaternary eolian and alluvial response to paleoclimate, Canyonlands, southeastern Utah: Geological Society of America Bulletin, v. 117, p. 1051–1069.

Rhodes, E.J., 2011, Optically Stimulated Luminescence Dating of Sediments over the Past 200,000 Years: Annual Review of Earth and Planetary Sciences, v. 39, p. 461–488, doi: 10.1146/annurev-earth-040610-133425.

Richmond, G.M., 1962, Quaternary stratigraphy of the La Sal Mountains, Utah: U.S. Geological Survey Professional Paper 324, 135 p., http://pubs.er.usgs.gov/publication/pp324 (accessed October 2017).

Rittenour, T.M., 2008, Luminescence dating of fluvial deposits: applications to geomorphic, palaeoseismic and archaeological research: Boreas, v. 37, p. 613– 635, doi: 10.1111/j.1502-3885.2008.00056.x.

Ritter, D.F., 1967, Terrace Development along the Front of the Beartooth Mountains, Southern Montana: GSA Bulletin, v. 78, p. 467–484, doi: 10.1130/0016- 7606(1967)78[467:TDATFO]2.0.CO;2.

76 Rockwell, T.K., Keller, E.A., Clark, M.N., and Johnson, D.L., 1984, Chronology and Rates of Faulting of Ventura River Terraces, California: Geological Society of America Bulletin, v. 95, doi: 10.1130/0016- 7606(1984)95<1466:CAROFO>2.0.CO;2.

Roy, M., Jordan, T.H., and Pederson, J., 2009, Colorado Plateau magmatism and uplift by warming of heterogeneous lithosphere: Nature, v. 459, p. 978–982, doi: 10.1038/nature08052.

Sinnock, S., 1981, Glacial moraines, terraces and pediments of Grand Valley, western Colorado, in Epis, R.C. and Callender, J.F. eds., Western Slope (Western Colorado), New Mexico Geological Society 32nd Annual Fall Field Conference Guidebook, p. 113–120.

Spencer, J.E., Peters, L., McIntosh, W.C., and Patchett, P.J., 2001, 40Ar/39Ar geochronology of the Hualapai Limestone and Bouse Formation and implications for the age of the lower Colorado River, in Young, R.A. and Spamer, E.E. eds., Colorado River: Origin and evolution, Grand Canyon, AZ, Grand Canyon Association Monograph 12, p. 89–91.

Stevenson, G.M., and Baars, D.L., 1986, The Paradox - A pull-apart basin of Pennsylvanian age, in Peterson, J.A. ed., Paleotectonics and sedimentation, American Association of Petroleum Geologists Memoir 41, p. 513–539, http://archives.datapages.com/data/specpubs/structu1/data/a155/a155/0001/0500/0 513.htm (accessed November 2017).

Sumsion, C.T., 1971, Geology and water resources of the Spanish Valley area, Grand and San Juan Counties, Utah: Utah Department of Natural Resources, Division of Water Rights Technical Publication 32, 45 p., https://pubs.er.usgs.gov/publication/70046223 (accessed January 2018).

Trudgill, B.D., 2011, Evolution of salt structures in the northern Paradox Basin: controls on evaporite deposition, salt wall growth and supra-salt stratigraphic architecture: Basin Research, v. 23, p. 208–238, doi: 10.1111/j.1365-2117.2010.00478.x.

U.S. Geological Survey, and Utah Geological Survey, 2006, Quaternary fault and fold database for the United States: https://earthquake.usgs.gov/hazards/qfaults/ (accessed November 2017).

Wagner, J.D.M., Cole, J.E., Beck, J.W., Patchett, P.J., Henderson, G.M., and Barnett, H.R., 2010, Moisture variability in the southwestern United States linked to abrupt glacial climate change: Nature Geoscience, v. 3, p. 110, doi: 10.1038/ngeo707.

Wallinga, J., Murray, A., and Wintle, A., 2000, The single-aliquot regenerative-dose (SAR) protocol applied to coarse-grain feldspar: Radiation Measurements, v. 32, p. 529–533, doi: 10.1016/S1350-4487(00)00091-3.

77 Wegmann, K.W., and Pazzaglia, F.J., 2002, Holocene strath terraces, climate change, and active tectonics: The Clearwater River basin, Olympic Peninsula, Washington State: GSA Bulletin, v. 114, p. 731–744, doi: 10.1130/0016- 7606(2002)114<0731:HSTCCA>2.0.CO;2.

Weisheit, J.S., and Fields, S.M., 2006, The Moab Mill Project: A technical report towards reclaiming uranium mill tailings along the Colorado River in Grand County, Utah: Living Rivers, 45 p. van Wijk, J.W., Baldridge, W.S., Hunen, J. van, Goes, S., Aster, R., Coblentz, D.D., Grand, S.P., and Ni, J., 2010, Small-scale convection at the edge of the Colorado Plateau: Implications for topography, magmatism, and evolution of Proterozoic lithosphere: Geology, v. 38, p. 611–614, doi: 10.1130/G31031.1.

York, D., Evensen, N.M., Martinez, M.L., and Delgado, J.D.B., 2004, Unified equations for the slope, intercept, and standard errors of the best straight line: American Journal of Physics, v. 72, p. 367–375.

Zaprowski, B.J., Evenson, E.B., Pazzaglia, F.J., and Epstein, J.B., 2001, Knickzone propagation in the Black Hills and northern High Plains: A different perspective on the late Cenozoic exhumation of the Laramide Rocky Mountains: Geology, v. 29, p. 547–550.

78 CHAPTER 3

TOPOGRAPHIC INSIGHTS INTO SALT TECTONICS AND INCISION

AROUND MOAB AND SPANISH VALLEY, UTAH

ABSTRACT

Rivers play a central role in the evolution of topography and are responsive to baselevel signals and substrate erodibility. This work uses several approaches in longitudinal profile analysis and paleotopographic reconstruction to document patterns in these drivers. Study drainages are tributaries of the Colorado River near the Moab-

Spanish Valley graben in east-central Utah, an actively-subsiding salt structure in the ancestral Paradox Basin of the central Colorado Plateau. Primary goals are to assess whether the transient incision recognized throughout the Colorado River system has propagated across the salt graben and up tributaries and whether baselevel signals are decipherable from the profiles of these bedrock streams. Results demonstrate that traditional long-profile and slope-area analysis cannot easily distinguish baselevel signals in this setting. Instead, patterns of concavity and steepness highlight the prominent lithologic control of this landscape. Knickpoints across the study area reflect both bedrock controls and transient incision, but are unable to independently resolve baselevel signals. Yet, transient baselevel signals are identified in Chi-transformed profiles, which are similar for drainages inside and outside the graben. Chi profiles suggest that tributaries are responding to a wave of middle-to-late-Pleistocene incision from the regional baselevel of the Colorado River and that any local baselevel signal from graben subsidence is obscured. The wave of incision has apparently propagated across the

79 subsiding lower reaches of the graben and it is currently distributed in upper bedrock

reaches of these drainages. Paleotopography reconstructions and chronostratigraphy

indicate that most graben subsidence has occurred since ~1.5 Ma, and the Pleistocene

pulse of regional incision is inferred to have coincided with an acceleration in salt

tectonism. This supports a hypothesized link between Colorado River incision and

subsurface salt-dissolution via groundwater infiltration. Though complicated by substrate

controls, integrating multiple long-profile analyses with an understanding of the geologic

template allows for tracing baselevel signals in bedrock dominated, erosional landscapes

like the Colorado Plateau.

INTRODUCTION

Using modern topography to understand landscape evolution is a fundamental

enterprise of tectonic geomorphology. River longitudinal profiles in particular are

sensitive to tectonic forcing, and they are commonly used to infer disequilibrium

processes such as differential uplift or transient incision (e.g. Wobus et al., 2006; Kirby

and Whipple, 2012). However, these interpretations are complicated in bedrock-

dominated landscapes with detachment-limited channels where variable strength exerts a

primary control. In these settings incision is not simply modulated by baselevel fall and

stream power. The Canyonlands region of the central Colorado Plateau in east-central

Utah is one such bedrock-controlled landscape, where the canyon-and-bench topography

is set by a succession of sub-horizontal strata of differing resistance. Understanding longitudinal profiles in this region is made even more challenging by the superimposed

80 effects of active salt tectonics, where local deformation and graben subsidence represent

potential local baselevel controls.

This work focuses on the high-relief region around Moab, Utah, where most

tributaries head in the tectonically-inactive La Sal Mountains laccolith (Fig. 3.1).

Drainages flow across upland mesas and through entrenched canyons of sub-horizontal

Mesozoic clastic sedimentary rocks. This famous lithologic sequence consists of friable sandstone and shale interbedded with well-indurated and silicified sandstone, as described in Chapter 4, Plate 2. Streams traversing this landscape can be characterized as detachment-limited bedrock systems, and provide the opportunity to test the competing effects of rock strength and baselevel fall on longitudinal profile. Farther downstream, drainages either cross the Moab-Spanish Valley graben or empty directly into the

Colorado River (Fig. 3.1). The alluvial Moab and Spanish valleys occupy the floor of the

~25 km long, northwest-trending graben, and are separated by a low bedrock saddle. The

Colorado River transects the northwest end of Moab Valley, and it is confined by

bedrock canyons both upstream and downstream. Analysis of tributaries that transect,

flow axially along, or do not encounter the Moab-Spanish Valley salt graben allows us to

test for potential baselevel signals from both the Colorado River and the subsiding

graben.

Past workers have hypothesized that subsidence in Moab-Spanish Valley is active

(Doelling et al., 2002), anomalously rapid (Guerrero et al., 2015), and has deformed

Quaternary alluvial deposits (Harden et al., 1985) based on mapping, paleoseismic

analysis, and soil correlations, respectively. Baselevel fall from this Quaternary

subsidence has left a poorly-understood imprint on the development of local drainages,

Figure 3.1. Map of the Moab-Spanish Valley watershed and study tributaries. Extent of the subsiding Moab/Spanish Valley salt graben is outlined in red. Black lines are previously-recognized Quaternary faults (U.S. Geological Survey and Utah Geological Survey, 2006). MF = Moab fault, KHFZ = Kayenta Heights fault zone as mapped by previous workers (Doelling, 2001; Doelling et al., 2002).

which are otherwise influenced by the region-wide incision of the Colorado River system. Specifically, there is evidence elsewhere on the Colorado Plateau for an upstream-migrating wave of incision hypothetically related to the initial integration of the

Colorado River through the Grand Canyon (Cook et al., 2009; Pederson et al., 2013b;

Darling and Whipple, 2015; Jochems and Pederson, 2015). It is unknown, however, whether such a signal has propagated intact across local salt structures and up minor tributaries to be recorded in the profiles of upstream detachment-limited reaches.

82 The goals of this research are to use digital analysis of longitudinal profiles and

relict topography to distinguish baselevel from lithologic signals, and to test for potential

transience in the local history of tributary incision and graben subsidence. Results are

interpreted in the context of the following questions: 1) What is the salt-tectonic and

erosional evolution of this famous landscape? 2) Can baselevel-fall signals be separated from rock-strength controls in this variable-bedrock, detachment-limited setting? 3) Do longitudinal profiles of study tributaries record a wave of transient Quaternary incision recognized elsewhere in the Colorado River system? 4) Does the subsiding graben impart on upstream tributaries an independent baselevel signal distinct from that of the Colorado

River? Results document stream profiles shaped dominantly but not exclusively by substrate resistance. These profiles encode subtle baselevel signals not detectable by all slope-area analyses, but they support a history of ongoing subsidence and transient incision linked to regional Colorado River incision. Results highlight the need for a holistic approach that accounts for the geologic template when applying longitudinal profile analysis to bedrock-dominated landscapes.

BACKGROUND

Cenozoic uplift and incision of the Colorado Plateau

The striking canyon topography of the Colorado Plateau is a result of Cenozoic

uplift and erosion. Epeirogenic uplift of the region began no later than 10 Ma (e.g Hunt,

1956; Morgan and Swanberg, 1985; Pederson et al., 2002), but the source of uplift is

heavily debated. Hypotheses include lithospheric hydration, heating, and buoyancy

related to removal of the Farallon slab in Laramide time (Humphreys et al., 2003; Roy et

83 al., 2009), asthenosphere convection and mantle drip over the past 10 Ma (Moucha et al.,

2009; van Wijk et al., 2010; Levander et al., 2011; Karlstrom et al., 2012), and isostatic rebound from the erosion and integration of the Colorado River system (Pederson et al.,

2002; Lazear et al., 2013). The ~2 km of uplift over the Cenozoic provided the potential for deep fluvial incision once external drainage was developed (Pederson et al., 2002).

It is widely accepted that the late-Cenozoic period of canyon cutting began with the integration of the Colorado River system through the Grand Canyon. This occurred sometime before 5.3 Ma (House et al., 2005; Dorsey et al., 2007; Karlstrom et al., 2014).

Drainage integration dropped the regional baselevel, necessarily initiating an upstream- propagating wave of incision through the Colorado Plateau. This period of Pliocene-

Pleistocene river incision is responsible for much of the Cenozoic exhumation of the central Colorado Plateau, as corroborated by thermochonologic data indicating a pointed acceleration in cooling rates beginning around 5 Ma (Hoffman et al., 2011; Murray et al.,

2016). Whether this pulse of incision is still migrating through the Colorado River system is incompletely understood, but several workers have documented evidence for transient incision in the central Colorado Plateau.

In tributaries of Cataract Canyon (Fig. 3.2) in the Needles Fault Zone, Geiger

(2014) used knickpoint elevations and relict profile projections to infer a pulse of incision currently stalled on resistant limestone beds. Using the elevation difference between projected paleo-profiles and modern tributary mouths together with interpolated incision rates (Pederson et al., 2013a), Geiger estimated that incision began propagating up tributaries from the Colorado River around 1.5 Ma. Salt deformation in the Needles Fault

Zone has potentially been active since middle Pleistocene time, and extensional spreading

84 and faulting along the basal salt contact is responsible for an unusual drainage organization where ephemeral washes terminate in closed-basin “swallow holes” (Geiger,

2014). It is notable that despite the confounding influences of non-uniform geology and active salt deformation, there is still a detectable transient incision signal in tributaries of

Cataract Canyon.

Figure 3.2. Regional map of the central Colorado Plateau and ancestral Paradox Basin. Canyons referenced in text are labeled. Salt graben are marked by white dashed lines, major folds are shown in black from Doelling (2004), and extent of evaporate facies according to Nuccio and Condon (1996) is marked by the black dashed line. Red box shows location of Figure 3.1.

Tributary profile analysis and cosmogenic nuclide dating of terraces in the Glen

Canyon region (Fig. 3.2) suggest an acceleration of incision from ~100 m/my to greater

85 than 300 m/my occurring sometime after 500 ka (Cook et al., 2009; Darling et al., 2012).

The timing of this acceleration post-dates the hypothesized incisional wave upstream in

Cataract Canyon by over one million years (Geiger, 2014), and thus may have been

triggered by a different episode of baselevel fall. These middle-to-late-Pleistocene incision rates in Glen Canyon are greater than those downstream in the Grand Canyon of

~140 m/my (Pederson et al., 2006), which is consistent with the recent passing of a wave of incision beyond the Grand Canyon. Cook et al. (2009) also modeled incision through upstream-dipping variable lithology, and suggest that the large knickpoint in the

Colorado River at Lee’s Ferry represents a transient incision signal that bifurcated at the upstream limit of resistant Paleozoic bedrock. The headward-migrating component of this signal may be responsible for the post-500 ka acceleration of incision in Glen Canyon, with the transient knickpoint at the head of this wave found somewhere upstream.

Jochems and Pederson (2015) document unexpectedly rapid late-Pleistocene incision rates, increasing upstream from 600 m/my to 900 m/my along a reach of the

Colorado River from Castle Creek to Dewey Bridge (30-50 km above Moab). These rates do not conform to previously hypothesized patterns of a “bullseye” of rapid incision in the central Colorado Plateau, which predicts incision rates of 300-400 m/my around

Moab (Pederson et al., 2013a). Instead, Jochems and Pederson (2015) suggest the high incision rates at Dewey Bridge record the same young wave of incision, which may be currently stalled in the Westwater and Desolation canyon knickzones (Fig. 3.2). It is unknown, however, whether such youthful transient incision is necessarily related to the integration of the Colorado River, or if there is a yet-unrecognized source of more recent baselevel fall that is responsible.

86 Salt deformation in the central Colorado Plateau

Localized deformation and dissolution of Pennsylvanian evaporites in the

ancestral Paradox Basin is superimposed on the broader uplift and incision history of the

central Colorado Plateau (Fig. 3.2). Passive salt deformation from differential overburden

pressure began shortly after deposition and lasted into the Triassic. After a period of

quiescence in the remainder of the Mesozoic, Laramide shortening reactivated salt

structures and produced the northwest-trending Paradox fold and fault belt (Baars and

Doelling, 1987; Doelling et al., 1988). The late-Cenozoic exhumation of the Colorado

Plateau is hypothesized to have initiated a period of ongoing dissolution as meteoric

waters penetrated salt structures and groundwater flow transported dissolved salts into

river systems (Cater, 1970; Huntoon, 1982; Doelling et al., 1988). This led to the

development of the suite of modern day northwest-trending graben that occupy the crests

of breached anticlines. Potential Quaternary salt-dissolution subsidence is recognized

throughout the Paradox Basin (Dane, 1935; Cater, 1970; Colman, 1983; Doelling et al.,

1988; Hudec, 1995; Gutiérrez, 2004), including in the Moab-Spanish Valley salt graben

and the nearby Castle Valley and Cache Valley graben of Arches National Park (Fig. 3.1)

(Doelling et al., 2002; Guerrero et al., 2015).

Geomorphic responses to baselevel fall

Baselevel refers to lowest elevation to which a river can erode (Powell, 1875) and it exerts a primary control on erosion and the development of drainage networks. In an effort to clarify usage of the term, Davis (1902) made the distinction between “regional” baselevel and a “local” baselevel. Regional baselevel refers to the terminus of a drainage

87 network below which streams cannot erode at the large scale, whereas local baselevel is the limit of erosion for a given reach, such as a waterfall held up on resistant bedrock.

Baselevel fall can be driven by eustacy, drainage integration and capture, or tectonics

(Leopold and Bull, 1979). To understand the development of Colorado River tributaries in the Moab region, it may be necessary to consider two distinct sources of baselevel fall: the incising Colorado River system and potentially the local subsiding salt graben. Mill and Pack creeks are not incised in their terminal reach across lower Moab Valley as they join the Colorado River, and instead flow across a broad floodplain marsh. This suggests that the graben (a local baselevel) is subsiding more rapidly than the Colorado River (a regional baselevel) is incising (Doelling et al., 2002; Guerrero et al., 2015). It is unclear, however, how these baselevel signals are manifest in the bedrock reaches of tributaries upstream of the graben.

River longitudinal profiles and bedrock controls

Rivers serve as conduits that transmit signals between headwater regions and sediment sinks. External perturbations provoke adjustments in channel gradient, width, and/or sediment load, and measurements of these adjustments can inform broader geological interpretations (Wobus et al., 2006). In erosional landscapes, the character and steepness of river longitudinal profiles are particularly useful in gauging responses to uplift, baselevel fall, or climate change (Kirby and Whipple, 2012). Given enough time to reach equilibrium, a river will adjust its slope so that it can transport the incoming sediment load or incise bedrock to meet baselevel with the available discharge and channel dimensions. This statement of equilibrium describes the graded condition, which

88 is a foundational concept from the early days of geomorphology (e.g. Davis, 1902;

Mackin, 1948). Theoretically, an equilibrium longitudinal profile is one where stream

power – the product of discharge and slope – is uniform along a river’s length (Leopold

and Langbein, 1962). Because discharge typically increases downstream, graded rivers

are thus expected to produce a smoothly concave-up profile. This relationship accurately describes the longitudinal profile of most alluvial streams (Sinha and Parker, 1996), but only applies to bedrock streams that cut through regions of uniform uplift and lithology

(e.g. Mackin, 1948; Hack, 1957; Whipple and Tucker, 1999). In such uniform conditions, bedrock incision (E) is commonly represented by the stream power erosion law,

= , (1) 𝑚𝑚 𝑛𝑛 where K𝐸𝐸 is a 𝐾𝐾dimensional𝐴𝐴 𝑆𝑆 erodibility coefficient, A is contributing drainage area used as a proxy for discharge, S is channel gradient, and m and n are positive constants relating to the physics of various erosion processes (Howard and Kerby, 1983; Whipple and Tucker,

1999). This relation predicts an intuitive coupling between channel gradient and rock erodibility that is documented in modeling, lab, and field studies.

Where a river encounters stronger rock it must adjust its width and gradient in order to increase stream power and maintain connection with baselevel (Mackin, 1948;

Whipple and Tucker, 1999). Most commonly this involves an increase in gradient and a decrease in channel width (e.g. Goldrick and Bishop, 1995; Montgomery, 2004). Though it does not conform to the theoretical concave-up graded profile, a stream that steepens over resistant bedrock may in fact be in equilibrium, as such a profile can persist through geologic time (Miller, 1991). In the Colorado Plateau, variations in rock strength are believed to exert a dominant control on river profiles. Pederson and Tressler (2012) found

89 a positive linear relation between stream power and a discharge-adjusted normalized steepness index along the Green and Colorado rivers, which they inferred to represent a coupling between steep reaches and harder bedrock. Bursztyn et al. (2015) confirmed this hypothesis by quantifying rock strength in the same region, again finding a positive linear relation between tensile strength and river gradient. These studies support the interpretation that equilibrium adjustment to variable rock strength, rather than baselevel- driven factors such as differential uplift or transient incision, are the primary determinates of longitudinal profile form in the Colorado River system.

Topographic indices

For a river profile in steady state with a constant uplift rate and uniform erodibility, the stream power erosion law can be simplified such that channel gradient scales to drainage area (e.g. Hack, 1957; Flint, 1974; Howard and Kerby, 1983). A common expression of this scaling is known as Flint’s Law,

= , (2) −𝜃𝜃 𝑠𝑠 where S𝑆𝑆 is channel𝑘𝑘 𝐴𝐴 gradient, A is upstream contributing area, ks is a channel steepness

index, and θ is a concavity index equal to the ration m/n (Flint, 1974). Plotting channel

gradient as a function of drainage area on log-log axes allows for the calculation of ks and

θ from the y-intercept and slope of the best-fit line, respectively. While ks can be used to

compare channel steepness within reaches of a single drainage, its value is highly

dependent on drainage area and concavity, making it unsuitable for comparison between

drainages of different size and shape. Because it is often necessary to compare such

90 drainages to infer the underlying tectonics, it is useful to define Flint’s Law in terms of a

fixed reference concavity (θref) and to calculate a normalized steepness (ksn) such that:

= , (3) −𝜃𝜃𝑟𝑟𝑟𝑟𝑟𝑟 𝑠𝑠𝑠𝑠 (Snyder𝑆𝑆 et al.,𝑘𝑘 2000;𝐴𝐴 Wobus et al., 2006; Kirby and Whipple, 2012). Expressed this way,

ksn can be used to identify channel reaches that do not conform to the expected profile of

a graded river specified by the value of θref.

In the central Colorado Plateau, active salt-dissolution subsidence exerts a control on tributary profiles that can be analyzed using Flint’s law. In the Needles Fault Zone,

subsidence in lower and middle tributary reaches is responsible for concavities higher

than would be expected for catchments where rock strength increases and precipitation

decreases downstream (Geiger, 2014). Upstream from Moab, the middle reach of Onion

Creek has a low concavity due to the ongoing uplift of the Onion Creek diapir, and Castle

Creek displays a high concavity as it flows into a subsiding graben (Colman, 1983;

Jochems and Pederson, 2015). Meanwhile, the profile of the mainstem Colorado River is

interpreted to remain unaffected by active subsidence in Professor Valley as the river

locally aggrades to maintain its transport gradient with downstream reaches (Jochems and

Pederson, 2015). Finally, while tributary concavity can be tied to salt activity, it is less clear whether high ksn values in tributaries upstream from Moab reflect salt deformation or lithologic controls (Jochems and Pederson, 2015).

Reaches of anomalous steepness in the longitudinal profile are known as

knickzones, and the inflection point at the upstream end is known as a knickpoint.

Together with the reach above, these features form local convexities in the longitudinal

profile. Knickzones (and by extension knickpoints) may be formed by baselevel fall,

91 changes in climate or groundwater hydrology, or variations in substrate erodability

(Mackin, 1948; Hack, 1957). In the case of baselevel fall or climate change, a knickpoint

is transmitted upstream through the drainage network as downstream reaches are

integrated into the new tectonic or climate regime (e.g. Gilbert, 1895). The river profile

returns to an equilibrium state only when the advancing knickpoint has fully migrated to

channel heads – a process that may require thousands to millions of years (Loget and Van

Den Driessche, 2009; Anderson and Anderson, 2010). Knickzones can also form across

reaches of resistant bedrock as predicted by the stream power erosion law, in which case

they are an equilibrium expression of the longitudinal profile (e.g. Miller, 1991; Grimaud

et al., 2014). For example, Pederson and Tressler (2012) interpret three of the four major knickzones on the Green and Colorado rivers as pinned on harder rocks of the Paleozoic and Proterozoic sections. Lithology-controlled knickzones are expected to slowly lengthen or recline in-place rather than propagate upstream (Gardner, 1983; Frankel et al., 2007; Grimaud et al., 2016).

Understanding the origins of knickpoints in a drainage network is fundamental to unlocking the geologic meaning of river profiles, and there are several ways to distinguish the transient from substrate-controlled cases. For example, ksn should be

uniform above and below a substrate-controlled knickpoint if the upstream and downstream reaches share a common lithologic and tectonic regime – as would be the case where a dike crosscuts tectonically-inactive, homogenous country rock. In the case of a transient knickpoint, ksn will be lower above the knickpoint than below, indicating that the upper reach is buffered from accelerated incision downstream (Snyder et al.,

2000). The planform and elevation distribution of knickpoints is also useful in this

92 distinction. A transient signal propagating through a drainage network in uniform

bedrock is expected to migrate at a constant vertical rate governed by the stream power

erosion law (Whipple and Tucker, 1999; Niemann et al., 2001), meaning that tributary knickpoints should occupy similar elevations (Wobus et al., 2006). Conversely, if knickpoints are substrate-controlled they should more closely follow the distribution of lithologic units than an elevation gradient – a pattern that has been documented throughout the Colorado River Basin (Pederson and Tressler, 2012; Bursztyn et al.,

2015). For example, in the northeastern Colorado Plateau Berlin and Anderson (2007)

found that knickpoint elevations in tributary streams almost universally coincide with a

single resistant, <50 m thick strata, and that knickpoint elevations vary according to the

regional dip. Finally, in the case of differing tectonic regimes (for example, a basin

spanning regions of high and low uplift rates), knickpoints may correspond with mapped

structures in planform (Wobus et al., 2006; Kirby and Whipple, 2012).

While steepness analysis using the slope-area relation has certain utility for interpreting river profiles, it is fundamentally limited by the pixelized nature of digital elevation models (DEMs) from which the slope and area data are produced. In particular, artifacts inherent in elevation data lead to substantial scatter when calculating slope – the derivative of elevation – meaning that slope-area plots are subject to uncertainty (Wobus et al., 2006; Perron and Royden, 2013). An alternative method that eliminates reliance on slope is to use the Chi parameter (χ). Chi is a proxy for longitudinal position in the drainage network calculated by integrating the steady-state stream power equation normalized to a reference area (Perron and Royden, 2013). It serves as an alternative variable by which river profiles can be examined, where channel elevation can be plotted

93 as a function of Chi as opposed to stream-wise distance. Such profiles are known as Chi plots, with the slope of the function representing a steepness index ratio between uplift rate and substrate erodibility (Perron and Royden, 2013). Rivers in steady state will plot

as a line, and any tributaries that are also in steady state should be co-linear. Because of

this, non-equilibrium channels with transient knickpoints are easily recognized in Chi

plots by having breaks in slope (Perron and Royden, 2013). Similarly, because Chi is a representation of relative position in the drainage network independent of basin size, transient signals with a common origin should collapse to a common position on the Chi axis for all tributaries in a catchment. In this regard, deviations from the steady state condition and the propagation of transient signals are easier to distinguish on Chi plots than on longitudinal profiles (Perron and Royden, 2013). Chi can also be mapped in

planform across a landscape to reveal contrasts in adjustment to baselevel fall and to

predict migration of drainage divides (Willett et al., 2014). Use of the Chi parameter thus

allows for differentiation between drainages that respond to distinct tectonic and

baselevel controls.

METHODS

Longitudinal profiles

Topography was analyzed through metrics derived from digital elevation models

(DEMs) using ArcGIS 10.3.1, Microsoft Excel 2016, and MATLAB R2017a. A 10-meter

resolution DEM was mosaicked for the study area from the USGS National Elevation

Dataset (U.S. Geological Survey, 2017a). Six tributaries that enter the Colorado River

within or just outside of the Spanish/Moab valley salt graben were selected for

94 topographic analysis (Fig. 3.1). North and South Mill creeks join courses 7.5 km upstream of the Colorado River in a bedrock canyon just outside the northeast graben margin. Upper Pack Creek and Brumley Creek confluence in an alluvial valley 27.5 km upstream of the Colorado River and 4 km upstream of the southeast graben terminus. The mainstem Mill and Pack creeks meet 2.1 km upstream of the Colorado River in the city of Moab. Courthouse Wash, which empties into the Colorado River from the north side of the graben, and Grandstaff Canyon, which joins the Colorado River from the south 5 km upstream of Moab Valley, were also chosen for analysis.

The five graben-intersecting drainages have different orientations, which allows us to test if these geometrical differences affect the potential preservation and propagation of baselevel signals. South Mill Creek flows parallel to but outside of the graben for ~16 km before meeting North Mill Creek and cutting perpendicularly across the graben margin into Moab Valley, while North Mill Creek maintains a course at a higher-angle to the graben for its entire length. The forks of Pack Creek headwater in the

La Sal Mountains before joining and flowing axially along the length of Spanish and

Moab Valleys. Courthouse Wash is entrenched in a bedrock canyon until within 0.3 km of its mouth, making first contact with the Colorado River and the salt graben simultaneously. Meanwhile, Grandstaff Canyon empties directly into the Colorado River upstream from the graben, and thus offers a control case of a drainage not affected by salt-dissolution subsidence.

Watershed boundaries were delineated using the Hydrology Toolbox in ArcGIS and Python batch processing. A flow accumulation area of 105 m2 (1000 cells) was chosen as the threshold for channel initiation based on visual air photo inspection and

95 field verification. Longitudinal profiles were constructed along the trunk drainage -

defined as the longest path from mouth to headwaters - for each watershed. Elevation,

contributing area, and upstream distance from the mouth were extracted from the DEM,

flow accumulation raster, and trunk drainage shapefile, respectively, at equal vertical

intervals of 12.192 m (40 ft) (Appendix C, Python script). This sampling window was

chosen because USGS 10 m DEMs are interpolated from 7.5 minute topographic data

with 40 ft contour intervals, and preserving the original vertical resolution reduces scatter in longitudinal profiles and slope-area plots (Snyder et al., 2000; Wobus et al., 2006).

Longitudinal profiles were plotted in Microsoft Excel from the stream-wise distance and

elevation data. No smoothing was necessary because the vertical sampling method

alleviates interpolation artifacts. This was a marked improvement over test-profiles for all

six drainages generated by equal-interval distance sampling, which contained numerous

artifacts and required smoothing.

An additional two alternative methods for generating longitudinal profiles were tested and ultimately excluded. The first involved “burning” a vector stream layer from the USGS National Hydrography Dataset (NHD) into the watershed DEM to aid in flow

routing (Maune, 2001; Wechsler, 2007). Visual comparison of burned and raw-DEM-

generated drainage networks showed that the burning process did not substantially alter

the drainage paths. This was expected, as the NHD vector layer was originally generated

from 10 m elevation data of the same resolution as the DEM used in our flow

accumulation model (U.S. Geological Survey, 2017b). The second method entailed

surveying latitude, longitude, and elevation along a 9 km reach of South Mill Creek with

a mapping-grade Trimble GPS. Points were collected at a background interval of 50 m

96 and at any inflection points in the profile such as knickpoints and riffles. Once the points

were post-process differentially corrected, a longitudinal profile was plotted to compare

with a profile from the raw, vertically-sampled DEM. The GPS profile contained

numerous elevation artifacts in the narrow canyon of South Mill Creek due to poor

satellite visibility and multipath error. Though the GPS profile was collected at a high

spatial resolution, the high frequency of artifacts made it no easier to distinguish true

knickpoints compared to the lower-resolution DEM profile. Longitudinal profiles generated from the vertical-sampling of unburned DEMs were thus used for the

remainder of the analysis.

Normalized steepness and concavity indices

The relation between channel gradient and upstream contributing area was

examined in Excel in order to identify knickzones and quantify profile steepness and

concavity. Gradient was calculated across a ~49 m (5 data point) vertical window, and

plotted against contributing area on logarithmic axes. The data were regressed with a

power law function in order to determine the channel steepness (ks) and concavity (θ)

indices from the slope-area relation of equation 2 (Flint, 1974). The concavity from this

equation, hereafter referred to as the native concavity (θnative), is the slope of the

regression line in log-slope/log-area space while the steepness index is taken from the y- intercept (Wobus et al., 2006). Because the six study tributaries have different drainage areas and native concavities, it was necessary to set a reference concavity (θref) to

calculate the normalized steepness index (ksn) (Snyder et al., 2000; Wobus et al., 2006). A

reference concavity of 0.35 was used for our analyses to allow direct comparison with ksn

97 from other regional studies (Geiger, 2014; Jochems and Pederson, 2015). Normalized steepness was calculated in Excel with the equation,

= 10 ( ) ( ), (4) 𝑙𝑙𝑙𝑙𝑙𝑙��𝑆𝑆��+ 𝜃𝜃𝑟𝑟𝑟𝑟𝑟𝑟�𝑙𝑙𝑙𝑙𝑙𝑙��𝐴𝐴�� 𝑠𝑠𝑠𝑠 which is𝑘𝑘 mathematically equivalent to forcing a linear regression in log(S)/log(A) space with a slope of θref.

A cumulative slope-area regression was performed on each trunk tributary from the watershed outlet to the lower limit of glacial deposits in order to remove recent overprinting on the longitudinal profile and to isolate baselevel and lithologic signals that pre-date the Last Glacial Maximum. Brumley Creek and South Mill Creek are the only drainages with glacial deposits in the headwaters, and slope-area analysis was truncated at elevations of 2680 and 3075 m, respectively, based on 1:100,000 scale maps (Doelling,

2001, 2004). In additional to whole-tributary regressions, slope-area data were regressed for 2-3 individual channel reaches within each profile. Reach boundaries were set by inflection points in the longitudinal profile corresponding to prominent knickpoints or the top of broad knickzones, and were identified by breaks in the slope-area relation. While all data points between the watershed outlet and the lower glacial limit were included for the cumulative tributary regression, reach ksn and θnative indices exclude data from the convex reach-bounding knickzones. In addition to these major knickzones, intra-reach knickpoints were identified on slope-area plots and longitudinal profiles. All knickpoints and knickzones were then classified by origin according to geologic maps (Doelling,

2001, 2004) and field relations. Finally, their planform and elevation distributions were mapped to assess potential transience or lithologic control (Crosby and Whipple, 2006;

Wobus et al., 2006).

98 Normalized steepness values from upper reaches were used to project longitudinal

profiles downstream in order to assess baselevel fall at the tributary mouth. These upper

tributary profiles are hypothesized to preserve relict topography pre-dating graben

formation, and can thus inform the magnitude of graben subsidence. Projected profiles

were anchored at the upper extent of the drawdown reach above the knickzone, which is

marked by the inflection point in concavity rather than the point of maximum steepness

(Gardner, 1983; Darling and Whipple, 2015). The uppermost classified knickzone or

knickpoint not formed by a Quaternary deposit was selected as the anchor for all six

tributary profiles. Profiles were projected in a piecewise manner using the calculated

normalized steepness of the upper channel reach and the blanket 0.35 reference

concavity. While some workers recalculate a reference concavity as the average θnative

from upper reaches across the study area (Schoenbohm et al., 2004), the variability in

lithology, topographic relief, and native concavity among our tributaries’ upper reaches

suggests that using the same θref as for the cumulative ksn analysis provides a more

meaningful comparison between profile projections. A projected slope at each point

along the profile was calculated by solving the normalized form of Flint’s Law (equation

3), where contributing area is extracted from the flow accumulation raster. Elevations were then calculated from the equation

= + [ ( )], (5)

𝑖𝑖+1 𝑖𝑖 𝑖𝑖 𝑖𝑖+1 where z𝑧𝑧i is the elevation𝑧𝑧 𝑆𝑆 ∗ at𝑥𝑥 a −point,𝑥𝑥 zi+1 is the elevation of the next downstream point, S is

the slope between the two points, and ( ) is the distance along the stream

𝑖𝑖 𝑖𝑖+1 channel between the two points (Geiger,𝑥𝑥 2014)− 𝑥𝑥 . The original zi is the true elevation of the

anchor point at the upper end of the knickzone. Profiles were also projected upstream

99 along the reach immediately above the knickzone to assess how well the projection fits

the existing channel profile.

Finally, normalized steepness values along tributary trunk drainages were mapped

in planform using the free and open-source TopoToolbox in MATLAB (Schwanghart and

Scherler, 2014). ksn was calculated over 500 m reaches with a reference concavity of

0.35, and exported as vector data to ArcGIS for display over background topography.

Normalized steepness is calculated in TopoToolbox from log-bin averaging and a nonlinear least squares regression of slope-area data in log-log space (Schwanghart and

Scherler, 2014). These planform map ksn values differ numerically from reach and whole-

profile ksn calculated in Excel, and are thus only used for relative comparison and for

assessing the spatial distribution of channel steepness. Classified knickzones and

knickpoints were displayed with the ksn map and underlying geology to further visualize

potential transient or lithologic controls on profile steepness.

Chi analysis

The Chi parameter (χ) was calculated along tributary trunk streams using

TopoToolbox functions in MATLAB to test whether these drainages are in steady state

equilibrium, contain transient incision signals, and/or respond to different baselevels

(Perron and Royden, 2013). Courthouse Wash was excluded from this analysis because it

has an anomalously low-gradient longitudinal profile compared to the other study

tributaries due to its low-elevation headwaters on the west side of the Colorado River.

DEMs clipped to each watershed were imported from ArcGIS to MATLAB, sinks were

filled, and a profile of elevation as a function of Chi (a Chi plot) was generated for each

100 trunk stream using the ‘chiplot’ tool in TopoToolbox. This tool also outputs the m/n value

– an expression of concavity – that best linearizes the Chi plot, as determined by a linear

least squares regression (Schwanghart and Scherler, 2014). The average best-fit m/n value of the five trunk drainages was then used as a reference concavity to re-run the

‘chiplot’ analysis, which allowed comparison between drainages. Chi plots were also generated from watershed DEMs of Mill and Pack creek tributaries that were truncated at the graben margin. This facilitated comparison of just the upper bedrock reaches independent of alluvial reaches along the graben floor.

Chi values were also aggregated in 30 m lengths and mapped in planform to identify potential disequilibrium along drainage divides and predict divide migration

(Willett et al., 2014). Chi maps included the entire drainage network derived from the original 10 m DEM and subsequent flow accumulation raster, as a large sample size of minor tributary headwaters is necessary to detect spatial differences in Chi across divides.

To further visualize these spatial patterns, vector Chi values were smoothed in an 800 m radius circular window using the focal statistics tool in ArcGIS to create a Chi raster. A slope map of this raster was then calculated and smoothed in a 200 m radius circular window to identify regions with the highest Chi gradient, which should correspond to divides with the greatest disequilibrium in baselevel signals (Willett et al., 2014).

Reconstruction of relict topography

To quantify graben subsidence over the Quaternary, a spline surface

representative of paleotopography was produced in GIS and compared with the modern

topography. Anchor points were digitized on the piedmont surface on the southwest flank

101 of the La Sal Mountains, and elevations were extracted from the 10 m DEM. A spline

surface was interpolated from the digitized points and elevations using the ArcGIS 3D

Analyst Spline tool. This surface was extended to produce a 90 m elevation raster

representing a paleo-piedmont spanning the entire Mill/Pack creek watershed and the

Moab-Spanish Valley graben. The spline was generated using a tension fit, a weight of

0.01, and considered the surrounding 15 points in its interpolation. These inputs were

chosen based on a validation study of topographic interpolation methods (Bergonse and

Reis, 2015) and the parameters used for other spline reconstructions in the Colorado

Plateau (Pederson et al., 2002).

To account for the thickness of Quaternary valley-fill gravels in estimating the

magnitude of graben subsidence, another spline surface representing depth-to-bedrock in

Moab and Spanish valleys was calculated based on alluvium thicknesses from water

wells (Lowe et al., 2007). Graben-interior well locations and depth-to-bedrock were

digitized, and points of zero depth were generated at 200 m intervals along the perimeter

of the alluvium/bedrock contact at the edge of the valley (Doelling, 2001, 2004). Only

wells that completely penetrated alluvium to the bedrock basin floor were used, such that

none of the depths anchoring the spline are minimum estimates. The alluvium thickness

spline was constructed with the same parameter inputs as the paleo-piedmont projection.

The resulting raster was added to the original DEM to produce a bedrock elevation raster.

This was then subtracted from the paleo-piedmont projection raster to create a difference

map. The mean pixel value of this difference raster represents an estimate for the average

depth of surface lowering below the paleo-piedmont, and multiplying the average depth by the basin area approximates the volume of material removed. This value is similar but

102 not equivalent to the commonly-used RVA index, the ratio of a catchment’s volume to area (Frankel and Pazzaglia, 2005). Because the elevation difference between the reconstructed piedmont and the bedrock floor of the graben represents total surface lowering (which may have occurred by processes other than subsidence), the volumetric estimates of subsidence are maximum values.

Seven other topographic indices were calculated in ArcGIS from DEMs clipped to individual watersheds. These include mean elevation, total relief, hypsometric integral

(HI), true RVA, and mean, median, and standard deviation of slope. Results from these calculations are presented in Appendix C (Table C.1; Fig. C.1).

RESULTS

Longitudinal profiles and knickzones

Longitudinal profiles of the study tributaries do not conform to the theoretical smoothly-concave equilibrium profile, and instead contain multiple knickzones and knickpoints (Fig. 3.3). With the exception of Courthouse Wash, all profiles display a prominent upper convexity and a knickzone that is ~3/4 up the drainage. The profiles of

North Mill, South Mill, Brumley, and upper Pack creeks are broadly concave, reflecting these drainages’ transition from headwaters in the La Sal Mountains to lower reaches along the alluvial graben floor. Grandstaff Canyon is the shortest study tributary and exhibits a two-stage profile marked by a convex headwaters and concave lower reaches

(Fig. 3.3E). The abrupt transition from convex to concave occurs where the stream descends from upland benches and becomes entrenched in a bedrock canyon. Courthouse

Wash is the exception to the regional pattern. It is the longest drainage and spans less

103 than half the relief of the other catchments, as it heads on low mesas northwest of the

Colorado River rather than in the La Sal Mountains (Fig. 3.1). Its profile is less concave

and less irregular than the other drainages, has only subtle knickzones, and it is

significantly less steep.

Each tributary was broken into two or three reaches separated by prominent

knickpoints that mark breaks in the slope-area relation and that are visible in the longitudinal profiles (Fig. 3.3). These reach-bounding knickpoints are largely controlled by lithology (Table 3.1). All study tributaries with the exception of Grandstaff Canyon encounter the well-cemented sandstone from the Moab Member of the Curtis Formation, which in each case produces a sharp knickpoint with a major knickzone reach below, highlighted in red in Figure 3.3 (k-3, 7, 13, 18, and 25). The Moab Member knickzone in

North Mill Creek (k-3) is especially prominent in the longitudinal profile (Fig. 3.3A).

Resistant interbeds in the Kayenta Formation and Dakota Sandstone are interpreted to control the upper reach-bounding knickzones in Grandstaff Canyon and North Mill

Creek, respectively. The two exceptions to the predominant lithologic control of reach- bounding knickzones are in Courthouse Wash and Grandstaff Canyon. The upper

Courthouse Wash knickpoint occurs in the weak, homogenous Navajo Sandstone, and the lower Grandstaff Canyon knickpoint is in the massively-bedded Wingate Sandstone.

Because these knickpoints are not associated with stratigraphic contacts or obvious resistant interbeds, they cannot be explained by lithology.

An additional 17 intra-reach knickpoints were identified from slope-area plots and longitudinal profiles (Table 3.1). As with the reach-bounding knickpoints, these features are mostly (13 of 17) controlled by substrate erodibility in the form of resistant bedrock

104

Figure 3.3. continues

105

Figure 3.3. Longitudinal profiles and slope-area plots for study tributaries. Profiles are shown at 10x vertical exaggeration. Reach-bounding knickzones are highlighted in red. Numbers key to knickpoints in Table 3.1. Only one projected profile (upper Pack Creek) is shown for clarity, implications discussed in text. Geology is shown conceptually. Qgt = glacial till, Qap = piedmont gravel, Qms = landslide deposits, Km = Mancos Shale, Kd = Dakota Sandstone, Kbc = Burro Canyon Formation, Jm = Morrison Formation, Jctm = Moab Member of the Curtis Formation, Jes = Entrada Sandstone, Jn = Navajo Sandstone, Jk = Kayenta Formation, Jw = Wingate Sandstone. Slope-area plots show regressions and true concavity (θ) calculated by reach, while ksn is reported with a reference concavity (θref) of 0.35. Reach-regressions exclude convex knickzones and glaciated headwaters.

106 TABLE 3.1. KNICKPOINTS SUMMARY Distance from Colorado River Elevation (km) (m) Classification* Substrate† ID North Mill Creek 31.6§ 2517 B Kd 1 30.5 2404 Q Qms 2 26.8 2093 B Jctm 3 20.2 1651 T Jk 4

South Mill Creek 39.0 2859 Q Qcb 5 37.4 2675 Q Qms 6 32.7 2280 B Jctm 7 28.0 1823 B Jk 8 20.6 1592 T Jn 9 6.3 1294 T Jn 10

Pack/Brumley Creek 40.3 2995 Q Qgt 11 37.9 2705 B Jctm 12 33.9 2212 B Jctm 13 29.4 1895 B Jctm 14 9.0 1335 B Jn 15

Upper Pack Creek 38.5 2763 B Jctm 16 37.5 2609 Q Qms 17 36.2 2459 B Jctm 18 35.5 2313 B Jctm 19

Grandstaff Canyon 20.7 2136 B Jn 20 19.1 1981 B Jk 21 16.0 1660 T Jw 22 12.1 1548 T Jw 23

Courthouse Wash 43.0 1489 T Jn 24 19.1 1306 B Jctm 25 11.5 1264 B Jcd 26 *Abbreviations: B = bedrock, T = transient, Q = Quaternary deposit; classification explained in text. †Geology mapped by Doelling (2001 and 2004). Qms = landslide deposit, Qcb = bouldery colluvium, Qgt = glacial till, Kd = Dakota Sandstone, Jctm = Moab Member of the Curtis Formation, Jcd = Dewey Bridge Member of the Carmel Formation, Jn = Navajo Sandstone, Jk = Kayenta Formation, Jw = Wingate Sandstone. §Reach-bounding knickzones in bold.

107 or channels choked by coarse Quaternary deposits. In aggregate, 9 of the 26 knickpoints

and knickzones in the study area are controlled by the resistant Moab Member of the

Curtis Formation. This shallowly-dipping lithology anchors a prominent bench at

elevations between 2000 and 2100 m on the flanks of the La Sal Mountains and is

responsible for the largest knickzones in the forks of Mill and Pack creeks. With the

addition of coarse interbedded fluvial sandstones like the Dakota Sandstone and the

Kayenta Formation, resistant lithologic units produce over half (15) of all characterized

knickpoints and knickzones. Coarse-grained Quaternary deposits are responsible for 5 knickzones in the upper forks of Mill and Pack creeks. These deposits include glacial moraines, large mass-movement deposits, and boulder-rich periglacial colluvium as

mapped by Doelling (2001 and 2004). The remaining 6 knickpoints are interpreted as

transient, as their position cannot be explained by substrate controls. The lowest three

knickpoints in the forks of Mill Creek (k-4, 9, and 10) are notable examples of transients

that mark the beginnings of strath terraces that grow in height downstream (Chapter 2).

Normalized steepness and concavity

Whole-catchment native concavity varies between 0.31 (Courthouse) and 0.78

(South Mill) and averages 0.46, which is close to the global average suggested by

previous studies (Flint, 1974; Snyder et al., 2000; Wobus et al., 2006) (Table 3.2). At the

reach scale, all six tributaries have relatively high concavity (θnative > 0.65) in their lowest

reach. It is noteworthy that this trend is common among both drainages that empty

directly into the Colorado River from incised bedrock canyons (Courthouse Wash and

Grandstaff Canyon) and those that first traverse an alluvial valley (Mill and Pack creeks).

108 The concavity of upper reaches is more variable, ranging from 0.17 (South Mill) to 2.29

(Brumley). Concavity contrasts within drainages support the classification of knickpoints

(Table 3.1). For example, the upper knickpoint in Courthouse Wash (k-24) is in the

homogenous Navajo Sandstone and separates reaches of identical concavity (θnative =

0.44), which matches the patterns predicted for a transient knickpoint in uniform bedrock

(Snyder et al., 2000; Wobus et al., 2006).

Normalized steepness calculated with θref = 0.45 allows comparison with other

global studies normalized to that value, and demonstrate first-order that Mill and Pack

creeks have reaches of high gradient (ksn > 100) (Table 3.2). The remaining results will refer to ksn values calculated with θref = 0.35, which is a regional average used in the studies of Geiger (2014) and Jochems and Pederson (2015). Courthouse Wash and South

Mill Creek record the lowest (4.0) and highest (26.0) whole-tributary ksn values,

respectively, which appear to follow total basin relief (Table 3.2). The highest individual-

reach normalized steepness (ksn > 28) are found in the upper reaches of South Mill,

Brumley, and upper Pack creeks – all of which headwater on the highest peaks of the La

Sal Mountains. Among the forks of Mill and Pack creeks, ksn is less variable in the lower

reaches (range = 4.3) than upper reaches (range = 26.2), reflecting the shared alluvial

substrate in the lowermost courses of these drainages. Lower reach normalized steepness

in the alluvial Mill and Pack creeks is notably higher (ksn >17) than in the bedrock-

confined lower reach of Grandstaff Canyon (ksn =11.2).

The planform distribution of normalized steepness, knickzones, and knickpoints

reflects the influence of substrate resistance on the study area stream profiles (Fig. 3.4).

Normalized steepness index values in this analysis come from TopoToolbox and are not

109 TABLE 3.2. CONCAVITY, STEEPNESS, AND RELIEF STATISTICS

ksn, θref = 0.35 ksn, θref = 0.45 Along-channel relief 0.7 0.9 * Reach θnative (m ) (m ) (m) North Mill Creek upper 0.44 11.99 44.59 256 middle 0.40 19.58 91.96 354 lower 0.80 17.65 101.96 744 whole 0.32 18.62 93.41 1585

South Mill Creek upper 0.17 28.03 139.75 719† lower 1.46 21.85 132.87 914 whole 0.44 25.98 145.14 1865†

Pack/Brumley Creek upper 2.29 38.21 191.39 426† lower 0.76 18.04 107.87 914 whole 0.78 23.44 131.77 1463†

Pack/upper Pack Creek upper 0.29 28.71 118.09 902 lower 0.88 21.13 123.67 1122 whole 0.45 25.63 128.27 2231

Grandstaff Canyon headwaters§ -0.81 9.55 36.83 110 upper 1.63 14.10 58.09 98 middle 1.24 14.26 66.84 256 lower 0.69 11.23 62.49 316 whole 0.43 13.77 65.13 1023

Courthouse Wash upper 0.44 3.49 14.31 110 middle 0.44 3.66 21.31 110 lower 0.79 4.91 34.99 73 whole 0.31 4.01 21.85 402 *Reach relief does not sum to whole-catchment relief due to exclusion of knickzones in reach measurements. †Value excludes glaciated reach. §Convex headwaters reach, see Fig. 3.3E.

110 directly comparable to those in Table 3.2, but the relative differences in ksn is informative of spatial patterns. In the forks of Mill and Pack creeks, normalized steepness is high across the Moab Member benches (Fig. 3.4, bright green) and through upstream reaches choked by coarse landslide and moraine deposits, which corresponds to a clustering of knickpoints and knickzones. Meanwhile, normalized steepness values are halved as these drainages flow through the softer Navajo Sandstone and Glen Canyon Group. The high ksn values in lower Courthouse Wash are anomalous, and are the result of DEM artifacts in its narrow bedrock gorge.

Knickpoints occupy a range in elevations from ~1250 to ~3000 m (Fig. 3.5) with an even distribution and no definitive clustering. Bedrock-controlled knickpoints have an especially even distribution, occurring at elevations from 1264 m to 2763 m. Quaternary deposit knickzones occur only above an elevation of 2400 m, which is because glacial/periglacial deposits and large landslide deposits are only found only on the high slopes of the La Sal Mountains. Knickpoints classified as transient exhibit the expected clustering, occupying a narrow elevation range between ~1300 and 1700 m (Crosby and

Whipple, 2006; Wobus et al., 2006). These elevations correspond to bedrock canyons entrenched in the predominately-homogenous eolian sandstones of the Glen Canyon

Group.

Upper-reach projections of longitudinal profiles yield very scattered results (Table

3.3). Because of this, only one example projected profile is shown in Figure 3.3. Six of the nine profiles unrealistically project below the modern elevation of the Colorado

River, with the upper Brumley Creek projection terminating below sea level. There is great scatter even among the three profiles that project above the modern Colorado River,

Figure 3.4. Normalized channel steepness map (θref = 0.35) and knickpoint locations. Red circles denote knickpoints used as reach boundaries. Tributaries are color-coded by normalized steepness index along 500 m reaches, with the high ksn reaches shown in red and the low ksn reaches shown in blue. Geology mapped by Doelling (2001 and 2004), with faults shown in black. The bright green polygons show the resistant Moab Member of the Curtis Formation while the dark green areas are underlain by the more erodible Glen Canyon Group. MV = Moab Valley, SV = Spanish Valley, MF = Moab fault, KHF = Kayenta Heights fault zone. The Moab Member coincides with high ksn and forms knickzones. 111

112

3000 bedrock transient

2500 Q deposit

2000 elevation (m)

1500

1000 0 10 20 30 40 distance from Colorado River (km) Figure 3.5. Knickpoints plotted by elevation and classification. Bedrock-controlled knickpoints span the elevation range, whereas transients are all lower than 1700 m.

with heights from 32 m to 422 m. The widely varying projection heights reflect differences in upper-reach steepness. Drainages with steep upper reaches and shallow- gradient lower reaches (South Mill, Brumley, upper Pack) project below the Colorado

River, while the shallow-gradient upper reach of Courthouse Wash projects to a higher elevation.

Chi analysis

Chi analysis allows the further exploration of profile steepness and knickpoint patterns. Figure 3.6 is a cumulative Chi plot showing Chi versus elevation profiles for the five tributaries analyzed. The origin of the plot is set at each tributary confluence with the

Colorado River, and the Chi integration is performed upstream from this datum. Chi values were generated using a common m/n of 0.41, which is the average of the

TopoToolbox best-fit concavity among the five profiles. This allows profile comparison

113 TABLE 3.3. LONGITUDINAL PROFILE PROJECTION DATA Projected elevation at Elevation difference Colorado River with modern profile Projected reach (m) (m) North Mill Creek Upper 1629 422 Middle 1077 -130

South Mill Creek Upper 745 -462

Brumley Creek Upper -65 -1272

Upper Pack Creek Upper 614 -593

Grandstaff Canyon Upper 1131 -77 Middle 1107 -101

Courthouse Wash Upper 1245 38 Middle 1239 32

between drainages. An initial observation is that none of the profiles are straight lines and

instead contain one major knickzone, illustrating again how these drainages are not

smoothly concave. This corroborates the fact that normalized steepness is highly variable

throughout the network (Fig. 3.4). The forks of Mill and Pack creeks are collinear from

the Colorado River through an elevation of ~1680 m. The Grandstaff Canyon profile

diverges slightly from this trend, and is nonlinear above its lower reach-bounding knickpoint at ~1530 m. Another key observation from Fig. 3.6 is that each tributary’s largest knickzone, which occurs in the Moab Member for the forks of Mill and Pack creeks and in the Kayenta Formation for Grandstaff Canyon, plots in the relatively narrow window between 5000 m and 6600 m on the Chi axis. This means that these

114 knickzones have the same relative contributing area, and lie in similar positions in the drainage network.

Figure 3.6. Cumulative Chi plots relative to a Colorado River datum. m/n =0.41. Major knickzones formed in resistant lithology are highlighted in thick red lines, and correspond to k-3, 7, 13, 18, and 21 in Table 3.1. Faults mapped by Doelling (2001 and 2004) are shown by small black arrows and are depicted in map-view in Figure 3.4. Note that knickzones display moderate clustering on the χ axis.

A different pattern is apparent when Chi and elevation are plotted relative to a starting datum at the graben margin along Mill and Pack creeks instead of the Colorado

River (Fig. 3.7). To generate this plot, the Chi integration was performed upstream from the graben margin rather than the tributary confluence with the Colorado River. The lower bound for the forks of Pack Creek is in the southeast arm of Spanish Valley where the Qatg2 terrace drops below grade (Chapter 2; Chapter 4, Plate 1). The lower bound for the Mill Creek tributaries is set at the Kayenta Heights fault in eastern Moab Valley.

115 Because it does not intersect the Moab-Spanish Valley salt graben, Grandstaff Canyon is excluded from this plot. These Chi profiles are intended to test whether the graben, rather than the Colorado River, is the main baselevel control for Mill and Pack creeks. Under this scenario the lower reaches of the four tributaries are still linear, suggesting a steady state at their downstream ends. However, they are not collinear, with the Mill Creek tributaries displaying a lower steepness than Pack Creek. Furthermore, the Moab Member knickzones do not converge as well on the Chi axis as they do in Figure 3.6, occupying a

2400 m range in Chi values that is 50% wider than when the integration is performed upstream from the Colorado River (Fig. 3.6).

Regional Chi maps show that escarpments tend to have the highest Chi gradients

(Fig. 3.8). For example, there is a sharp gradient between low Chi in the valley floor and high Chi at the top of the escarpment along the southwest margin of Spanish Valley (Fig.

3.8B, point 1). Also of note is the contrast between low Chi in the lower reach of North

Mill Creek and high Chi in the neighboring tributaries of Grandstaff Canyon (Fig. 3.8B, point 2). A final example of Chi disequilibrium exists between South Mill Creek (lower-

Chi) and tributaries of Pack Creek (higher-Chi) in southeast Spanish Valley (Fig. 3.8B, point 3). The trunk course of South Mill Creek in this location is directly adjacent to first- order tributaries of Pack Creek, yet the channel sits 100 m higher than the floor of

Spanish Valley.

Paleotopography reconstruction

To estimate the magnitude of Quaternary subsidence within the Moab-Spanish

116

Figure 3.7. Cumulative Chi plots relative to a graben floor datum. m/n =0.41. Major knickzones formed in resistant lithology are highlighted in thick red lines, and correspond to k-3, 7, 13, and 18 in Table 3.1. Faults mapped by Doelling (2001 and 2004) are shown by small black arrows and are depicted in map-view in Figure 3.4. Arrow at origin is the Kayenta Heights fault zone at Mill Creek. Note that knickzones do not converge as well on χ axis as in Figure 3.6.

Valley graben, we reconstructed paleotopography anchored by high remnant piedmont gravels that were mapped by Doelling (2001 and 2004). Previous workers have hypothesized the piedmont to be early Pleistocene in age (Richmond, 1962; Harden et al.,

1985; Doelling, 2004). Two CRN geochronology samples indicate deposition wasoccurring at least between ~1.6 and ~1.7 Ma (Chapter 2). An estimated age of 1.5 Ma is considered to represent the landform top and the last period of aggradation in this gravel package, as the ~1.6 Ma CRN sample was collected from a depth of 17 m in the

117

Figure 3.8. Regional Chi maps. Extent is the same as Figure 3.1. (A) Vector Chi map of the regional drainage network. (B) Chi gradients. Numbers refer to examples discussed in the text. Catchments outlined in white: NM = North Mill, SM = South Mill, P = Pack, B = Brumley, G = Grandstaff, C = Courthouse.

118 deposit (Chapter 2). An elevation-difference map between the 1.5 Ma reconstructed topography and the bedrock floor of the graben was created using splines of the piedmont and of published alluvium thicknesses beneath Moab and Spanish valleys (Lowe et al.,

2007) (Fig. 3.9). Reported difference values (Table 3.4) correspond to total surface lowering, which may have occurred by processes other than salt dissolution alone. Thus, values in Table 3.4 offer maximum constraints on salt-dissolution subsidence.

The greatest surface lowering has occurred near the Colorado River in northwest

Moab Valley, and another lowering center is in east-central Spanish Valley (Fig. 3.9B).

Negative values in the difference raster are found along the southwest graben escarpment, which protrudes above the reconstructed paleo-piedmont and suggests there was positive relief to the southwest prior to initial subsidence of Moab and Spanish valleys. Maximum and spatially-averaged surface lowering magnitudes are taken from the maximum and mean raster values, respectively. The maximum surface lowering depth is 265 m in

Spanish Valley and 277 m in Moab Valley, while mean surface lowering is 133 ± 78 in

Spanish Valley and 145 ± 83 m in Moab Valley (Table 3.4). The large 1σ errors reflect the uncertainty in projecting the paleo-piedmont beyond the anchor points (Fig. 3.9) as well as the irregular bedrock topography beneath the valley-fill alluvium.

DISCUSSION

Geologic influences on longitudinal profile form

The Canyonlands region of the central Colorado Plateau is a famous bedrock landscape renowned for its scenic slickrock, slot canyons, towers, and escarpments.

Longitudinal profiles of drainages near Moab and Spanish valleys confirm the

119

Figure 3.9. Paleotopographic reconstruction of piedmonts above Moab (MV) and Spanish (SV) valleys. (A) Anchor points for piedmont (green) and alluvium (blue) splines. Piedmont extent (teal) from Doelling (2001 and 2004), alluvium depths from Lowe et al. (2007). (B) Difference raster between projected piedmont spline and the bedrock floor of the Moab-Spanish Valley graben. Red colors show subsidence, blues show positive relief. Cosmogenic burial samples (red stars) suggest abandonment age for piedmont surface of ~1.5 Ma. Note that piedmont was projected ~15 km northwest beyond anchor points.

120 TABLE 3.4. PALEOTOPOGRAPHIC DIFFERENCING STATISTICS Volume Surface lowering Area removed (m) Extent (km2) (km3) Mean* Maximum Moab 145 ± Valley 22.3 3.23 83 277 Spanish 133 ± Valley 55.4 7.38 78 265 Entire 130 ± graben 87.2 11.31 86 277 *Mean surface lowering reported with 1σ uncertainty.

fundamental role of lithology in shaping the topography of this region. Instead of conforming to the smoothly-concave profile expected under conditions of uniform bedrock and uplift, our tributary profiles contain numerous substrate-controlled knickpoints and knickzones.Because of this, the most conspicuous pattern of knickzone/knickpoint distribution in planform relates to the distribution of the resistant

Moab Member of the Curtis Formation along upland benches and Quaternary moraines and landslide deposits on the high flanks of the La Sal Mountains (Fig. 3.4).

Slope-area data also bear a signature of heterogeneous bedrock in the study area.

Reach-concavity is highly variable, and appears most closely related to the distribution of shallowly-dipping lithologic units in the bedrock segments of the profile. The juxtaposition of upstream-resistant and downstream-erodible lithology results in reaches with abnormally high concavities, with resistant units forming broad reach-bounding convexities. The lower reaches of North and South Mill creeks have high concavities

(θnative >0.75) well above the global average of 0.4-0.45 (Flint, 1974; Snyder et al., 2000;

Wobus et al., 2006). The resistant Moab Member forms knickzones at the head of these reaches, while the underlying soft and homogenous Navajo Sandstone permits streams to

121 quickly transition to a much lower gradient (Fig. 3.3A and B). Brumley and upper Pack

creeks exhibit the same pattern, with the contrast between the Moab Member on the

flanks of the La Sal Mountains and easily-erodible alluvium in Spanish Valley producing highly-concave lower reaches (Fig. 3.3C and D). The headwater reach of Brumley Creek

has the highest observed concavity in the study area (θnative = 2.29), which can be

explained by the fact that it flows though the soft shales of the Morrison Formation

before encountering the Moab Member bench at its lower-bounding knickzone (Fig.

3.3C). The upper reach of neighboring South Mill Creek is bound by a knickzone in the

same lithology, but its profile is much straighter (θnative = 0.17) because it crosses a planar

piedmont ramp choked with coarse landslide deposits immediately above the Moab

Member. This mutes the prominence of the lower-bounding knickzone in the profile and

results in a much lower concavity.

Normalized steepness is highly correlative to the geologic and topographic template, with large values highlighting areas of resistant substrate in headwater reaches.

For example, the farthest upstream reach of upper Pack Creek has a high normalized steepness (ksn = 28.7) due to encountering boulder-rich landslide deposits (Fig. 3.3D).

There is also an intuitive relation between catchment relief and normalized steepness

(Table 3.2). South Mill, upper Pack, and Brumley creeks headwater in the La Sal

Mountains and display the highest values (ksn >23), while Courthouse Wash flows from

the low plateaus northwest of the Colorado River and has a normalized steepness that is

much lower (ksn = 4). Interestingly, the lower alluvial reaches of Mill and Pack creeks

have higher normalized steepness (ksn > 17) than the lower bedrock-confined reach of

Grandstaff Canyon (ksn = 11.2). This suggests either that the coarse porphyry bedload in

122 Moab and Spanish valleys is more resistant than the Glen Canyon Group sandstones in

Grandstaff Canyon or that the greater overall relief of the Mill and Pack catchments is

expressed as steeper stream gradients even in the lowest reaches.

Despite their clear influence, geologic and substrate controls cannot explain all

patterns documented in longitudinal profiles and slope-area analysis. Six knickpoints are found in homogenous bedrock far from lithologic contacts and are interpreted as transient. These features display the vertical-step morphology and pattern of emanating strath terraces that is characteristic of transient knickpoints (Gardner, 1983; Wobus et al.,

2006; Pazzaglia, 2013). Theoretical slope-area relations of transient knickpoints (Snyder et al., 2000; Wobus et al., 2006) are observed for the upper Courthouse Wash knickpoint

(k-24), which separates reaches of identical concavity and upstream-decreasing ksn (Fig.

3.3F). Aside from Courthouse Wash, all classified transient knickpoints occur in the

lower halves of the drainage network and are below the most prominent bedrock-

controlled knickzones (Figs. 3.3 and 3.4).

Chi analysis: Implications for regional transience and baselevel fall

Though partly obscured by bedrock controls in slope-area and knickpoint

distribution analyses, baselevel signals and transient incision are manifest in Chi analysis.

With Chi and elevation calculated relative to a Colorado River datum, the largest

knickzones in Grandstaff Canyon and the forks of Mill and Pack creeks (k-3, 7, 13, 18,

21) collapse to a narrow ~1600 m window on the Chi axis (Fig. 3.6). This means that

despite their different elevations, catchment sizes, and orientations relative to the graben,

these knickzones occupy comparable relative positions in the watershed according to

123 their contributing area. In turn, this suggests that these large knickzones are related to a

common baselevel origin and likely represent a propagating wave of incision (Perron and

Royden, 2013).

Given a hypothesis of transient incision, Chi profiles also inform where that

incision originated. The Chi transform extending upstream from the Colorado River

datum shows the Mill, Pack, and Grandstaff drainages to have collinear lower reaches

(Fig. 3.6). The slope of a Chi profile is a steepness index, and collinear profiles are

diagnostic of drainages that are in equilibrium with the same ratio of baselevel fall (U) to

substrate resistance (K) (Perron and Royden, 2013). The Chi transform does not

distinguish which component plays a greater role in establishing this equilibrium, so we

cannot say whether baselevel or substrate resistance is more responsible for the

collinearity of lower reaches in Fig. 3.6. However, the combination of these factors

(expressed as the ratio U/K) must be equal among the drainages, either because baselevel

fall and substrate resistance are in fact equivalent across the drainage network or because

drainages with weaker substrate happen to be compensated by experiencing a less-rapid history of baselevel fall. We favor the former case, and suggest that the cobble-rich porphyry bedload in the lower alluvial reaches of Mill and Pack creeks along the graben floor may in fact have a similar K-value to the Glen Canyon Group sandstones that

Grandstaff Canyon cuts through in its incised lower bedrock reach.

The lower reaches of Mill and Pack creeks are not collinear when the Chi transform extends upstream from a graben margin origin (Fig. 3.7). This means that Mill and Pack creeks have differing U/K ratios relative to the floor of the salt graben, but equal U/K ratios relative to the Colorado River (Fig. 3.6). Put another way, these

124 drainages share an equilibrium adjustment with baselevel fall from the Colorado River but not the graben floor. This implies that if there are indeed shared baselevel signals among these drainages, they originate from the incising Colorado River rather than the subsiding salt graben. The collinearity of lower Grandstaff Canyon (below k-23), along with the collapse of its knickzone in Chi space with Mill and Pack creeks, supports this interpretation (Fig. 3.6). Grandstaff Canyon meets the Colorado River 5 km upstream from Moab Valley, and its watershed does not intersect the Moab-Spanish Valley graben.

It is possible that subsidence in Moab Valley creates a baselevel fall signal that propagates up the Colorado River and into Grandstaff Canyon. However, it is more likely that internal sedimentation feedbacks in Moab Valley mask any local baselevel signal of graben subsidence given that the Colorado River maintains its transport gradient across the alluvial valley between the bedrock reaches upstream and downstream of Moab. We thus favor the interpretation that the baselevel signal shared between Grandstaff Canyon,

Mill Creek, and Pack Creek originates from the Colorado River and that the upper reaches of these drainages are responding to regional incision.

Reconciling transient incision and bedrock control

The interpretation of knickzones and knickpoints around Moab and Spanish valleys, and indeed in any non-equilibrium landscapes with variable bedrock, is more nuanced than the typical endmembers of fixed (lithology control) versus transient

(baselevel control). Instead, we document knickzones that are expressions of a transient wave of incision originating from the Colorado River that is hung up and distributed onto the most resistant lithologies. Similar cases of transient incision stalling and splitting in

125 resistant bedrock have been interpreted elsewhere in the Colorado Plateau in relation to

Colorado River integration (Cook et al., 2009; Pelletier, 2010; Darling et al., 2012). It is

possible then that the wave of incision around Moab and Spanish valleys has propagated

at slightly different rates through each tributary system due to variations in substrate, but

that in each case it has slowed and focused upon the resistant units it encounters. The

large Moab Member knickzones are the most prominent expression of the baselevel fall

signal that is now distributed throughout the Mill/Pack drainage network. This signal

must also be present in smaller knickpoints both upstream and downstream that have

formed on slightly less abrupt lithologic discontinuities or in homogenous bedrock (Fig.

3.4).

Not all knickpoints are directly related to baselevel fall from the Colorado River, as there are additional controls. The lowest transient knickpoints in Mill Creek (k-4, 9, and 10) are interpreted to have retreated from the Kayenta Heights fault zone (KHFZ), which forms the northeast margin of Moab Valley (Chapter 4, Plate 3). For example, k-

10 occurs in the weak, homogenous Navajo Sandstone yet is expressed as a prominent waterfall (Chapter 2). This requires an episode of local baselevel fall, most readily from movement along the KHFZ only 1 km downstream, to explain. However, inasmuch as subsidence of the Moab-Spanish Valley graben may ultimately keep pace with Colorado

River incision through groundwater infiltration and dissolution (Chapter 2), downstream regional baselevel fall may be a driver of graben fault slip and can thus be considered an indirect control on the lower Mill Creek transient knickpoints. The upper knickpoint in

Courthouse Wash could relate to the same wave of incision responsible for the Moab

Member knickzones in Mill and Pack creeks, with its anomalously high position in the

126 drainage network explained by rapid propagation in a low-relief catchment (Table 3.2).

Alternatively, this knickpoint could relate to local baselevel fall from early-Quaternary movement along the nearby Moab fault (Fig. 3.4), which others have deemed possible based on scarp-retreat rates (Olig et al., 1996).

Despite these local baselevel controls related to potentially complex feedbacks between incision and salt dissolution, Chi analysis demonstrates that regional baselevel fall along the Colorado River, which is manifest in the largest knickzones in the

Mill/Pack and Grandstaff watersheds, is responsible for the vast majority of transient incision in this landscape. Chi profiles alone, however, cannot constrain the timing of this incision episode. It is possible that the pulse of incision is related to the ~5 Ma integration of the Colorado River, which is a widely-recognized source of baselevel fall in the central

Colorado Plateau (e.g. Pederson et al., 2002). Alternatively, the transient wave may be related to a more recent acceleration of incision of unknown source recognized in the

Glen Canyon region at ~500 ka (Cook et al., 2009; Darling et al., 2012). Terrace chronostratigraphy in Mill and Pack creeks record a major acceleration of incision from

<40 m/my (0.1 m/ky) to >600 m/my (0.6 m/ky) sometime before 200 ka in the middle

Pleistocene (Chapter 2). Given constraints on the timing of initial graben subsidence

(presented below), a prior-to-200 ka wave of incision would have encountered and propagated through the actively-subsiding Moab-Spanish Valley graben in a structural setting similar to the modern day. It is possible that earlier signals of baselevel fall and transient incision propagated through Colorado River tributary precursors to Mill and

Pack creeks before the Moab salt anticline was breached, but the records of these episodes are no longer prominent in stream profiles or terrace deposits. Instead, the

127 middle-to-late-Pleistocene wave of incision has overprinted any earlier signals and has

been a key influence in the modern canyon topography of this region.

Paleotopographic and geochronologic constraints on graben evolution

The ~1.6 Ma cosmogenic burial date of the Johnson’s Ridge gravel deposit

(Chapter 2) has important implications for the initiation of graben subsidence. This preserved upland gravel represents the past position of a drainage that transported sediment from the La Sal Mountains. It is reasonable to assume that this paleo-drainage would seek the lowest position in the landscape over which to flow, so there could not have been a large valley to the southwest at the time that the drainage deposited the

Johnson’s Ridge gravel (~1.6 Ma). This suggests that the 240 m of basin-margin relief in modern times between Johnson’s Ridge and the bottom of Spanish Valley near Ken’s

Lake must have been produced after ~1.5 Ma, the estimated abandonment age of the upland gravel tread. Because there are no higher remnant gravels or escarpments on the northeast side of Spanish Valley, there is no evidence for faulting and graben subsidence pre-dating the Johnson’s Ridge gravels. Based on exposures that show imbrication to the west-northwest (not parallel to the modern South Mill Creek), as well as the presence of analogous features today along the north, east, and south flanks of the La Sal Mountains

(Doelling, 2001, 2004), the Johnson’s Ridge gravels are interpreted as discrete remnants of a once-extensive piedmont bajada.

An alternate hypothesis is that the Johnson’s Ridge gravels represent shoestring deposits from South Mill Creek that record its ~1.5 Ma confinement in a bedrock canyon comparable to today. This is inspired by the striking similarity in trend and width of the

128 Johnson’s Ridge gravels with the modern canyon-bottom of South Mill Creek and the

contrast in plan view geometry between the Johnson’s Ridge and age-equivalent South

Mountain upland gravels (Chapter 4, Plate 1). The landform of Johnson’s Ridge is ~3 km

long and <300 m wide for an aspect ratio of >10:1 (Chapter 4). The farthest northwest

equivalent Johnson’s Ridge upland gravel deposit is ~20 km from the La Sal Mountain

headwaters (Chapter 2; Chapter 4), which is 7 km more distant than the preserved toe of

the South Mountain piedmont south of the study area. These deposits are further

anomalous in that they are preserved in only a narrow, linear zone and are absent from

the broad bench topography between North and South Mill creeks, a distribution which is

also mirrored by that of the younger T6 deposits of Johnson’s Up On Top (Chapter 2;

Chapter 4, Plate 1). Following the shoestring hypothesis, initial graben collapse lowered

the local baselevel of South Mill Creek near the Colorado River, triggering headward

incision, but was not extensive enough to capture and divert the drainage into the graben

interior. After abandonment of the Johnson’s Ridge surface, a combination of graben

subsidence to the southwest and fluvial erosion to the northeast would have eroded the

bedrock paleo-canyon walls, resulting in a unique case of inverted topography where

unconsolidated channel deposits were stranded atop a drainage divide. Under such a

scenario the Pleistocene relief production along the northeast graben margin would have

been even greater than the present-day 240 m of relief between Johnson’s Ridge and the

floor of Spanish Valley, as there would have been a higher and no-longer-preserved

bedrock plateau confining South Mill Creek to the southwest around ~1.6 Ma. However,

given the balance of field evidence for analogous modern piedmonts on the flanks of the

La Sal Mountains and clast imbrication that is not parallel to the modern course of South

129 Mill Creek, the paleo-piedmont hypothesis is preferred over the alternate shoestring

gravel hypothesis.

A GIS projection of the Johnson’s Ridge and South Mountain paleo-piedmont

informs what the landscape may have looked like before subsidence of the Moab-Spanish

Valley salt graben initiated. This projected piedmont surface is 277 m above the lowest

point in the bedrock floor of Moab Valley, and the mean depth beneath this surface to the

modern bedrock floor of the graben is 130 ± 86 m. The mean and maximum surface

lowering magnitudes are within 8% when Moab and Spanish valleys are considered

separately (Table 3.4). This similarity, which belies well data showing much deeper

Quaternary alluvium in Moab Valley than Spanish Valley (Lowe et al., 2007), likely

reflects the slope of the paleo-piedmont away from the La Sal Mountains such that its

tread was much lower over present-day Moab Valley than present-day Spanish Valley in

the early Pleistocene.

When considered in light of late-Pleistocene subsidence rates calculated from chronostratigraphy (Chapter 2), these depths of surface lowering – which are maximum estimates of salt-driven subsidence – imply that subsidence must have accelerated over the Pleistocene. Based on chronostratigraphy and geometric projections of deformed terraces, subsidence since ~200 ka has occurred at ~450 m/my (0.45 m/ky) in Spanish

Valley and at least ~440 m/my (0.44 m/ky) in Moab Valley (Chapter 2). At its minimum subsidence rate (which is taken from the slip rate calculated across the Kayenta Heights fault zone) the maximum amount of surface lowering in Moab Valley could have been accomplished in just ~640 ky. Yet cosmogenic radionuclide dating of the Johnson’s

Ridge and South Mountain piedmonts estimates that at least ~1.5 my elapsed between the

130 abandonment of these broad surfaces (which may have marked initiation of graben

subsidence) and the present day. Thus, the late-Pleistocene subsidence rates calculated

for Moab and Spanish valleys cannot have been sustained throughout the entire

Pleistocene, suggesting an acceleration of subsidence at some time in the Pleistocene. It

is possible that this acceleration occurred in parallel and was linked to the middle-to-late-

Pleistocene wave of incision recorded in local Chi profiles and the terrace

chronostratigraphy of upper South Mill Creek (Chapter 2). Considering the hypothesized

connection between groundwater infiltration and salt-dissolution (Huntoon, 1982;

Gutiérrez, 2004; Jochems and Pederson, 2015), it is intuitive to expect a pulse of river incision and baselevel lowering to be accompanied by a corresponding pulse of subsidence.

The paleo-piedmont reconstruction illustrates the likelihood of an escarpment at

~1.5 Ma along the southwest margin of Moab and Spanish valleys analogous to the

modern-day Moab Rim. It is probable that the escarpment was farther northeast at ~1.5

Ma than it is today, and has since retreated to the southwest with the initiation of graben

subsidence. This ~1.5 Ma escarpment would have also had less topographic relief than in

modern times, with the toe of the projected paleo-piedmont extending to its base (Fig.

3.9). Regardless, rather than indicating an early expression of the salt graben, this

escarpment is taken to represent paleo-relief on the southern section of the Moab fault similar to that seen today northwest of Moab Valley (Olig et al., 1996; Foxford et al.,

1998). Pre-graben deformation on the Moab fault is considered distinct from later

dissolution-induced subsidence accommodated across the same fault zone (Olig et al.,

1996; Black et al., 2004; Chapter 4). Thus, this paleo-relief along the southwest

131 escarpment does not contradict the interpretation that graben collapse has only occurred in the last 1.5 million years (Chapter 2). In other words, the paleo-piedmont reconstruction leads to a similar interpretation for the evolution of the Moab-Spanish

Valley salt graben as implied by the distribution and cosmogenic burial ages of upland gravels.

A baselevel fall history may also be informed by relict topography preserved above major tributary knickzones. To test this, we projected upper reaches using a common reference concavity (θref = 0.35) and the normalized steepness unique to each reach. It was hypothesized that measuring the elevation difference between the mouth of the modern-day tributary and the projected relict profile would inform patterns of baselevel fall along the Colorado River. However, the results instead demonstrate the influence of substrate resistance on the normalized steepness of upper reaches, which obscures any baselevel signal that relict profiles may contain. For example, upper Pack

Creek cuts through boulder-rich landslide deposits and resistant bedrock, highlighted by the Moab Member of the Curtis Formation, which yields a steep channel with a relatively high ksn of ~29 (Fig. 3.3D). This high steepness produces a profile that projects almost

600 m below the modern-day elevation of the Colorado River, which is an unrealistic paleo-mouth elevation in this clearly erosional landscape. There is substantial scatter among paleo-profiles, and most project below the elevation of the Colorado River (Table

3.3). These observations of substrate-overprinting on profile projections discourage us from making any baselevel interpretations from this analysis.

Topography along the graben margin provides insights into the evolution of faulting and subsidence in this system. Chi values are relatively low in Spanish Valley

132 and high on the bedrock rims just outside the graben (Fig. 3.8A). This sharp low-to-high

Chi gradient is best seen along the southwest escarpment of Spanish Valley (Fig. 3.8B,

point 1), and suggests that drainage divides should be migrating outward from the graben

axis (Willett et al., 2014). This prediction is consistent with patterns of graben fault-zone

evolution that have been hypothesized from fault-trench geochronology, which suggest

that the youngest, most active faults are the ones most distal to the graben axis (Guerrero

et al., 2015). Consequently, we hypothesize that the Moab-Spanish Valley graben has been progressively widening though time, and will continue to do so through cliff retreat that is most rapid along the southwest escarpment.

Summary

The central Colorado Plateau is a famous bedrock landscape that records signs of transient landscape evolution. In tributaries of the Moab-Spanish Valley salt graben, incision signals from the Colorado River overshadow the local baselevel control of graben subsidence. It is apparent that regional baselevel fall in the Colorado River system has initiated a migrating wave of incision that is currently focused and distributed onto resistant lithologic units in the middle and upper reaches of these tributaries. Based on terrace chronostratigraphy (Chapter 2), it is interpreted that this pulse of incision reached local tributaries prior to 200 ka in the middle Pleistocene. Graben subsidence may have accelerated in parallel with incision, and in the late Pleistocene these two processes have occurred at similar rates (Chapter 2). This supports the hypothesized process link between Colorado River incision and subsurface salt-dissolution through groundwater infiltration. Based on subsidence rates and the distribution of upland gravel deposits, we

133 suggest that salt-dissolution collapse did not begin until after ~1.5 Ma, and has since produced all of the topographic relief along the northeast graben margin. More broadly, this work illustrates that it is possible to extract baselevel signals from stream profiles in

a bedrock-dominated landscape by combining integrative (Chi) and derivative (slope- area) analyses with a holistic understanding of the geologic template.

REFERENCES

Anderson, R.S., and Anderson, S.P., 2010, Geomorphology: The Mechanics and Chemistry of Landscapes: Cambridge University Press, 655 p.

Baars, D.L., and Doelling, H.H., 1987, Moab salt-intruded anticline, east-central Utah, in Beus, S.S., ed., Rocky Mountain Section of the Geological Society of America, Centennial Field Guide, Boulder, CO, v. 2, p. 275–280.

Bergonse, R., and Reis, E., 2015, Reconstructing pre-erosion topography using spatial interpolation techniques: A validation-based approach: Journal of Geographical Sciences, v. 25, p. 196–210, doi: 10.1007/s11442-015-1162-2.

Berlin, M.M., and Anderson, R.S., 2007, Modeling of knickpoint retreat on the Roan Plateau, western Colorado: Journal of Geophysical Research: Earth Surface, v. 112, p. F03S06, doi: 10.1029/2006JF000553.

Cater, F.W., 1970, Geology of the salt anticline region in southwestern Colorado: U.S. Geological Survey Professional Paper 637, 80 p., https://pubs.usgs.gov/pp/0637/report.pdf.

Colman, S.M., 1983, Influence of the Onion Creek salt diapir on the late Cenozoic history of Fisher Valley, southeastern Utah: Geology, v. 11, p. 240–243, doi: 10.1130/0091-7613(1983)11<240:IOTOCS>2.0.CO;2.

134 Cook, K.L., Whipple, K.X., Heimsath, A.M., and Hanks, T.C., 2009, Rapid incision of the Colorado River in Glen Canyon – insights from channel profiles, local incision rates, and modeling of lithologic controls: Earth Surface Processes and Landforms, v. 34, p. 994–1010, doi: 10.1002/esp.1790.

Crosby, B.T., and Whipple, K.X., 2006, Knickpoint initiation and distribution within fluvial networks: 236 waterfalls in the Waipaoa River, North Island, New Zealand: Geomorphology, v. 82, p. 16–38, doi: 10.1016/j.geomorph.2005.08.023.

Dane, C.H., 1935, Geology of the Salt Valley anticline and adjacent areas, Grand County, Utah: USGS Numbered Series 863, 184 p., https://pubs.er.usgs.gov/publication/b863.

Darling, A., and Whipple, K., 2015, Geomorphic constraints on the age of the western Grand Canyon: Geosphere, v. 11, p. 958–976, doi: 10.1130/GES01131.1.

Davis, W.M., 1902, Baselevel, Grade and Peneplain: The Journal of Geology, v. 10, p. 77–111.

Doelling, H.H., 2004, Geologic Map of the La Sal 30’ x 60’ Quadrangle, San Juan, Wayne, and Garfield Counties, Utah, and Montrose and San Miguel Counties, Colorado: Utah Geological Survey.

Doelling, H.H., 2001, Geologic map of the Moab and eastern part of the San Rafael Desert 30’ x 60’ quadrangles, Grand and Emery Counties, Utah, and Mesa County, Colorado: Utah Geological Survey.

Doelling, H.H., Oviatt, C.G., and Huntoon, P.W., 1988, Salt deformation in the Paradox region: Utah Geological Survey, Bulletin 122.

Doelling, H.H., Ross, M.L., and Mulvey, W.E., 2002, Geologic Map of the Moab 7.5’ Quadrangle, Grand County, Utah: Utah Geological Survey.

Flint, J.J., 1974, Stream gradient as a function of order, magnitude, and discharge: Water Resources Research, v. 10, p. 969–973, doi: 10.1029/WR010i005p00969.

135 Foxford, K.A., Walsh, J.J., Watterson, J., Garden, I.R., Guscott, S.C., and Burley, S.D., 1998, Structure and content of the Moab Fault Zone, Utah, USA, and its implications for fault seal prediction: Geological Society, London, Special Publications, v. 147, p. 87–103, doi: 10.1144/GSL.SP.1998.147.01.06.

Frankel, K.L., and Pazzaglia, F., 2005, Tectonic geomorphology, drainage basin metrics, and active mountain fronts: Geografia Fisica e Dinamica Quaternaria, v. 28, p. 7– 21.

Frankel, K.L., Pazzaglia, F.J., and Vaughn, J.D., 2007, Knickpoint evolution in a vertically bedded substrate, upstream-dipping terraces, and Atlantic slope bedrock channels: GSA Bulletin, v. 119, p. 476–486, doi: 10.1130/B25965.1.

Gardner, T.W., 1983, Experimental study of knickpoint and longitudinal profile evolution in cohesive, homogeneous material: GSA Bulletin, v. 94, p. 664–672, doi: 10.1130/0016-7606(1983)94<664:ESOKAL>2.0.CO;2.

Gilbert, G.K., 1895, Niagara Falls and their history: New York, American Book Company, The Physiography of the United States, http://archive.org/details/cu31924028832363 (accessed November 2017).

Goldrick, G., and Bishop, P., 1995, Differentiating the roles of lithology and uplift in the steepening of bedrock river long profiles: an example from southeastern Australia: The Journal of Geology, v. 103, p. 227–231, doi: 10.1086/629738.

Grimaud, J.-L., Chardon, D., and Beauvais, A., 2014, Very long-term incision dynamics of big rivers: Earth and Planetary Science Letters, v. 405, p. 74–84, doi: 10.1016/j.epsl.2014.08.021.

Grimaud, J.-L., Paola, C., and Voller, V., 2016, Experimental migration of knickpoints: influence of style of base-level fall and bed lithology: Earth Surface Dynamics, v. 4, p. 11–23, doi: https://doi.org/10.5194/esurf-4-11-2016.

136 Hack, J.T., 1957, Studies of longitudinal stream profiles in Virginia and Maryland.: U.S. Geological Survey Professional Paper 294-B, 97 p., https://pubs.usgs.gov/pp/0294b/report.pdf (accessed September 2017).

Harden, D.R., Biggar, N.E., and Gillam, M.L., 1985, Quaternary deposits and soils in and around Spanish Valley, Utah: GSA Special Papers, v. 203, p. 43–64, doi: https://dx.doi.org/10.1130/SPE203.

Hoffman, M., Stockli, D.F., Kelley, S.A., Pederson, J., and Lee, J., 2011, Mio-Pliocene erosional exhumation of the central Colorado Plateau, eastern Utah - New insights from apatitie (U-Th)/He thermochronometry, in CREvolution 2 - Origin and Evolution of the Colorado River System, Workshop Abstracts, U.S. Geological Survey Open-File Report 2011-1210, p. 132–136.

House, P.K., Pearthree, P.A., Howard, K.A., Bell, J.W., Perkins, M.E., Faulds, J.E., and Brock, A.L., 2005, Birth of the lower Colorado River—Stratigraphic and geomorphic evidence for its inception near the conjunction of Nevada, Arizona, and California, in Pederson, J.L. and Dehler, C.M. eds., Interior Western United States, Geological Society of America Field Guide 6, p. 357–387, doi: 10.1130/2005.fld006(17).

Howard, A.D., and Kerby, G., 1983, Channel changes in badlands: GSA Bulletin, v. 94, p. 739–752, doi: 10.1130/0016-7606(1983)94<739:CCIB>2.0.CO;2.

Hudec, M.R., 1995, The Onion Creek salt diapir: an exposed diapir fall structure in the Paradox Basin, Utah, in Salt, Sediment and Hydrocarbons; Papers Presented at the Gulf Coast Section Society of Economic Paleontologists and Mineralogists Foundation Sixteenth Annual Research Conference, v. 16, p. 125–134.

Hunt, C.B., 1956, Cenozoic geology of the Colorado Plateau: U.S. Geological Survey Professional Paper 279, 99 p., http://pubs.er.usgs.gov/publication/pp279 (accessed October 2017).

Jochems, A.P., and Pederson, J.L., 2015, Active salt deformation and rapid, transient incision along the Colorado River near Moab, Utah: Journal of Geophysical Research: Earth Surface, v. 120, p. 2014JF003169, doi: 10.1002/2014JF003169.

137 Karlstrom, K.E., Coblentz, D., Dueker, K., Ouimet, W., Kirby, E., Wijk, J.V., Schmandt, B., Kelley, S., Lazear, G., Crossey, L.J., Crow, R., Aslan, A., Darling, A., Aster, R., et al., 2012, Mantle-driven dynamic uplift of the Rocky Mountains and Colorado Plateau and its surface response: Toward a unified hypothesis: Lithosphere, v. 4, p. 3–22, doi: 10.1130/L150.1.

Karlstrom, K.E., Lee, J.P., Kelley, S.A., Crow, R.S., Crossey, L.J., Young, R.A., Lazear, G., Beard, L.S., Ricketts, J.W., Fox, M., and Shuster, D.L., 2014, Formation of the Grand Canyon 5 to 6 million years ago through integration of older palaeocanyons: Nature Geoscience, v. 7, p. ngeo2065, doi: 10.1038/ngeo2065.

Kirby, E., and Whipple, K.X., 2012, Expression of active tectonics in erosional landscapes: Journal of Structural Geology, v. 44, p. 54–75, doi: 10.1016/j.jsg.2012.07.009.

Leopold, L.B., and Bull, W.B., 1979, Base level, aggradation, and grade: Proceedings of the American Philosophical Society, v. 123, p. 168–202.

Leopold, L.B., and Langbein, W.B., 1962, The concept of entropy in landscape evolution: U.S. Geological Survey Professional Paper 500-A, 21 p., http://pubs.er.usgs.gov/publication/pp500A (accessed October 2017).

Loget, N., and Van Den Driessche, J., 2009, Wave train model for knickpoint migration: Geomorphology, v. 106, p. 376–382, doi: 10.1016/j.geomorph.2008.10.017.

Mackin, J.H., 1948, Concept of the graded river: GSA Bulletin, v. 59, p. 463–512, doi: 10.1130/0016-7606(1948)59[463:COTGR]2.0.CO;2.

Maune, D.F. (Ed.), 2001, Digital Elevation Model Technologies and Applications: The DEM Users Manual: The American Society for Photogrammetry and Remote Sensing.

138 Miller, J.R., 1991, The influence of bedrock geology on knickpoint development and channel-bed degradation along downcutting streams in south-central Indiana: The Journal of Geology, v. 99, p. 591–605.

Montgomery, D.R., 2004, Observations on the role of lithology in strath terrace formation and bedrock channel width: American Journal of Science, v. 304, p. 454–476, doi: 10.2475/ajs.304.5.454.

Morgan, P., and Swanberg, C.A., 1985, On the Cenozoic uplift and tectonic stability of the Colorado Plateau: Journal of Geodynamics, v. 3, p. 39–63, doi: 10.1016/0264- 3707(85)90021-3.

Niemann, J.D., Gasparini, N.M., Tucker, G.E., and Bras, R.L., 2001, A quantitative evaluation of Playfair’s law and its use in testing long-term stream erosion models: Earth Surface Processes and Landforms, v. 26, p. 1317–1332, doi: 10.1002/esp.272.

Olig, S.S., Fenton, C.H., McCleary, J., and Wong, I.G., 1996, The Earthquake Potential of the Moab Fault and Its Relation to Salt Tectonics in the Paradox Basin, Utah, in Huffman, A.C.J., Lund, W.R., and Godwin, L.H. eds., Geology and Resources of the Paradox Basin, Salt Lake City, UT, Utah Geological Association Guidebook 25, p. 251–264, http://archives.datapages.com/data/uga/data/067/067001/251_ugs670251.htm (accessed November 2017).

Pazzaglia, F.J., 2013, 9.22 Fluvial terraces, in Shroder, J.F. ed., Treatise on Geomorphology, San Diego, CA, Academic Press, p. 379–412.

Pederson, J.L., Anders, M.D., Rittenhour, T.M., Sharp, W.D., Gosse, J.C., and Karlstrom, K.E., 2006, Using fill terraces to understand incision rates and evolution of the Colorado River in eastern Grand Canyon, Arizona: Journal of Geophysical Research: Earth Surface, v. 111, p. F02003, doi: 10.1029/2004JF000201.

139 Pederson, J.L., Burnside, N., Shipton, Z., and Rittenour, T., 2013b, Rapid river incision across an inactive fault—Implications for patterns of erosion and deformation in the central Colorado Plateau: Lithosphere, v. 5, p. 513–520, doi: 10.1130/L282.1.

Pederson, J.L., Mackley, R.D., and Eddleman, James L., 2002, Colorado Plateau uplift and erosion evaluated using GIS: GSA Today, v. 12, p. 4–10.

Pederson, J.L., and Tressler, C., 2012, Colorado River long-profile metrics, knickzones and their meaning: Earth and Planetary Science Letters, v. 345–348, p. 171–179, doi: 10.1016/j.epsl.2012.06.047.

Pelletier, J.D., 2010, Numerical modeling of the late Cenozoic geomorphic evolution of Grand Canyon, Arizona: GSA Bulletin, v. 122, p. 595–608, doi: 10.1130/B26403.1.

Perron, J.T., and Royden, L., 2013, An integral approach to bedrock river profile analysis: Earth Surface Processes and Landforms, v. 38, p. 570–576, doi: 10.1002/esp.3302.

Powell, J.W., 1875, Exploration of the Colorado River of the West and its tributaries, 1869-1872: Washington, Government Printing Office, Smithsonian Institute Annual Report, 291 p.

Richmond, G.M., 1962, Quaternary stratigraphy of the La Sal Mountains, Utah: U.S. Geological Survey Professional Paper 324, 135 p., http://pubs.er.usgs.gov/publication/pp324 (accessed October 2017).

Roy, M., Jordan, T.H., and Pederson, J., 2009, Colorado Plateau magmatism and uplift by warming of heterogeneous lithosphere: Nature, v. 459, p. 978–982, doi: 10.1038/nature08052.

Schoenbohm, L.M., Whipple, K.X., Burchfiel, B.C., and Chen, L., 2004, Geomorphic constraints on surface uplift, exhumation, and plateau growth in the Red River region, Yunnan Province, China: Geological Society of America Bulletin, v. 116, p. 895, doi: 10.1130/B25364.1.

Schwanghart, W., and Scherler, D., 2014, Short Communication: TopoToolbox 2 – MATLAB-based software for topographic analysis and modeling in Earth surface sciences: Earth Surface Dynamics, v. 2, p. 1–7, doi: https://doi.org/10.5194/esurf- 2-1-2014.

140 Sinha, S.K., and Parker, G., 1996, Causes of Concavity in Longitudinal Profiles of Rivers: Water Resources Research, v. 32, p. 1417–1428, doi: 10.1029/95WR03819.

Snyder, N.P., Whipple, K.X., Tucker, G., and Merritts, D., 2000, Landscape response to tectonic forcing: DEM analysis of stream profiles in the Mendocino Triple Junction region, northern California: Geological Society of America Bulletin, v. 112, p. 1250–1263.

U.S. Geological Survey, 2017a, 10 & 30 Meter Elevation Models (USGS NED): Utah GIS Portal, https://gis.utah.gov/data/elevation-terrain-data/10-30-meter-elevation- models-usgs-ned/ (accessed September 2017).

U.S. Geological Survey, 2017b, U.S. Geological Survey - National Hydrography Dataset:, https://nhd.usgs.gov/ (accessed September 2017).

U.S. Geological Survey, and Utah Geological Survey, 2006, Quaternary fault and fold database for the United States:, https://earthquake.usgs.gov/hazards/qfaults/ (accessed November 2017).

Wechsler, S.P., 2007, Uncertainties associated with digital elevation models for hydrologic applications: a review: Hydrology and Earth System Science, v. 11, p. 1481–1500, doi: 10.5194/hess-11-1481-2007.

Whipple, K.X., and Tucker, G.E., 1999, Dynamics of the stream-power river incision model: Implications for height limits of mountain ranges, landscape response timescales, and research needs: Journal of Geophysical Research: Solid Earth, v. 104, p. 17661–17674, doi: 10.1029/1999JB900120. van Wijk, J.W., Baldridge, W.S., Hunen, J. van, Goes, S., Aster, R., Coblentz, D.D., Grand, S.P., and Ni, J., 2010, Small-scale convection at the edge of the Colorado Plateau: Implications for topography, magmatism, and evolution of Proterozoic lithosphere: Geology, v. 38, p. 611–614, doi: 10.1130/G31031.1.

Willett, S.D., McCoy, S.W., Perron, J.T., Goren, L., and Chen, C.Y., 2014, Dynamic Reorganization of River Basins: Science, v. 343, p. 1248765, doi: 10.1126/science.1248765.

Wobus, C., X Whipple, K., Kirby, E., Snyder, N., Johnson, J., Spyropolou, K., Crosby, B., and Sheehan, D., 2006, Tectonics from topography: Procedures, promise, and pitfalls: Geological Society of America Special Paper, v. 398, p. 55–742398, doi: 10.1130/2006.2398(04).

141 CHAPTER 4

GEOLOGIC MAP OF THE SOUTHERN HALF OF THE RILL CREEK

AND NORTHERN HALF OF THE KANE SPRINGS

QUADRANGLES, UTAH

ABSTRACT

The adjoining halves of the Rill Creek and Kane Springs 7.5’ quadrangles are east

of Moab, Utah. This area includes the eastern half of the Moab-Spanish Valley salt

graben and the neighboring bedrock plateaus to the southwest and northeast. Mapping of

this quadrangle-sized region is part of a broader effort to understand active salt

deformation, and the associated landscape evolution and geologic hazards in the ancestral

Paradox Basin. Strata from early Jurassic to late Cretaceous in age are exposed in the

map area, and Quaternary units include alluvial, colluvial, eolian, mass-wasting, and

fluvial terrace deposits. Graben subsidence is accommodated by systems of shallowly-

seated, near-vertical, gravitational faults along the margins of Spanish Valley. The two graben margin fault zones display contrasting deformation styles and fault geometries.

Ongoing Quaternary subsidence in Spanish Valley is documented in the spatial and temporal distribution of terrace deposits along Mill and Pack creeks (Chapter 2), which

confirms previous hypothesis of active salt deformation. The hazard of active, non-

seismic, salt-dissolution collapse and faulting appears to be modest, with greater concern relating to attendant mass-wasting processes along the valley margins.

142 INTRODUCTION

The map area spans the southern half of the Rill Creek and northern half of the

Kane Springs 7.5’ quadrangles, which are in eastern Utah in Grand and San Juan counties, east of the city of Moab (Fig. 4.1). Much of the map area encompasses the eastern part of the Moab-Spanish Valley salt graben, which is ~25 km long, 2.5 to 5 km wide, and trends northwest-southeast. The graben contains two alluvial valleys – Moab

Valley and Spanish Valley – separated by a low topographic saddle of shallowly-covered bedrock. The map area encompasses the entirety of Spanish Valley, and the eastern border intersects the bedrock saddle. The region outside of the Moab-Spanish Valley salt graben is characterized by erosional topography typical of the Colorado Plateau, with bedrock benches and uplands bound by escarpments and dissected by drainage networks.

The neighboring Moab 7.5’ quadrangle was mapped by Doelling et al. (2002). The Kane

Springs quadrangle, previously known as the Mount Peale-4 quadrangle, was mapped photogeologically by Hackman (1956), while the Rill Creek quadrangle has not been previously mapped at 1:24,000 scale. The Rill Creek and Kane Springs quadrangles were included in the 1:100,000 scale map compilations by Doelling (2001 and 2004). In terms of Quaternary geology, Richmond (1962) completed initial mapping and unit description in the La Sal Mountains and Spanish Valley. Harden et al. (1985) proposed correlations and ages of Spanish Valley alluvial deposits based on soil carbonate concentrations, and

Guerrero et al. (2015) mapped graben-wide terrace gravel deposits as part of a paleoseismic study. Past workers have hypothesized that subsidence of the Moab-Spanish

143 Valley graben is active, locally rapid, and is marked by deformed Quaternary alluvial deposits (Harden et al., 1985; Doelling et al., 2002; Guerrero et al., 2015).

Figure 4.1. Map area for the southern half of the Rill Creek and the northern half of the Kane Springs 7.5’ quadrangles. Map boundary is outlined in black. Red lines are Quaternary faults (U.S. Geological Survey and Utah Geological Survey, 2006). Gray rectangle on inset map shows location within the state of Utah.

Mapping the distribution of Quaternary deposits allows testing of previous hypotheses of active subsidence, and documenting the geometry of graben faults contributes understanding to the mechanics of graben collapse. Mapping is also motivated by geohazard concerns related to salt-dissolution faulting and subsidence around the community of Moab, which have gained attention in the wake of sinkhole openings beneath a subdivision in 2015 (Willis, personal communication). The city of

144 Moab is an important economic hub for eastern Utah, relying primarily on outdoor

recreation tourism as well as mineral extraction and agriculture. A growing residential

population and increased tourism have pushed development southeast into Spanish

Valley, spurring the need for this geologic mapping.

Field mapping was initially conducted at 1:8,000 scale in Spanish Valley, along

the graben margins, and in North Mill, South Mill, and Pack creeks. Field lines were then

digitized using VrTwo photogrammetry software. The bedrock plateau southwest of

Spanish Valley, the uplands between the forks of Mill Creek, and Wilson and South

Mesas were initially mapped in VrTwo and subsequently field-verified. VrTwo lines

were imported into ArcGIS, where the final map layout was created. Plate 1 is the

1:24,000 scale map. Unit descriptions, a cross section, and Quaternary correlations are on

Plate 2. A map extension from the western edge of the Rill Creek quadrangle to the graben margin along Mill Creek is included as Plate 3 due to this area’s importance in the

authors’ related work. Mapping was conducted in collaboration with the Utah Geological

Survey (UGS) and partially funded by the USGS EDMAP program, award

G16AC00205.

SETTING

Geographic setting

The map area includes Spanish Valley, which is the southeast extent of the Moab-

Spanish Valley salt graben, and dissected bedrock plateaus to the northeast and southwest

(Fig. 4.1). Drainages above the southwest rim of Spanish Valley flow away from the

graben, feeding Hunter’s Canyon, Kane Creek, and eventually the Colorado River. To the

145 northeast of Spanish Valley a lower-elevation bedrock plateau is dissected by the forks of

Mill Creek. Farther east an escarpment bounds Wilson and South mesas, which form a prominent bench anchored by resistant lithology along the eastern map boundary. Outside the map area to the east, the laccolithic La Sal Mountains form the headwaters for the three primary drainages that traverse the Rill Creek and Kane Springs quadrangles (Fig.

4.1). North and South Mill creek are entrenched in bedrock canyons for their entire lengths in the map area, but eventually converge and enter Moab Valley orthogonally from the east. Pack Creek enters the graben axially near its southeast terminus and flows through Spanish Valley before exiting the map area, becoming entrenched across the bedrock saddle, and joining Mill Creek in Moab Valley. Mainstem Mill Creek then joins the Colorado River just prior to the river’s reentrance into a bedrock canyon after transecting Moab Valley. Mill and Pack creeks are perennial streams, but are heavily diverted for agricultural and domestic use. Elevations in the map area range from 1320 m along Mill Creek to 2200m on South Mesa.

Geologic setting

The Moab-Spanish Valley salt graben is in an area of salt-deformation associated with the ancestral Paradox Basin of the central Colorado Plateau (Fig. 4.2). The Paradox

Basin was a flexural-subsidence foreland basin along the southwestern flank of the basement-cored Uncompahgre Uplift that developed in middle Pennsylvanian time

(Stevenson and Baars, 1986; Barbeau, 2003). The restricted basin was periodically connected with a sea to the west, and up to 2500 m of evaporites precipitated from hypersaline brines during times of basin isolation (Trudgill, 2011). These deposits

146 comprise the Paradox Formation, a sequence of 33 evaporite cycles of anhydrite, dolomite, organic-rich calcareous shale, and halite (Raup and Hite, 1992). Salt deposition spanned an area stretching roughly 325 km northwest-southeast and 150 km northeast- southwest in modern-day southeast Utah and southwest Colorado (Barbeau, 2003) (Fig.

4.2). Deposition of the Paradox Formation ceased by the end of the Pennsylvanian, though subsidence in the Paradox Basin continued into the late Triassic (Baars and

Doelling, 1987; Doelling et al., 1988).

Figure 4.2. Regional map of the ancestral Paradox Basin. Salt graben are marked by white dashed lines, and major folds are shown in black. Red box shows location of Figure 4.1. Paradox Basin evaporite facies extent from Nuccio and Condon (1996), structures from Doelling (2004).

147 Ductile deformation from differential overburden pressure began during deposition, as evaporites migrated laterally and rose buoyantly along pre-existing faults.

The resulting diapirs and salt walls grew to up to 4500 m vertically, and deformed the overlying strata into a series of anticlines and synclines (Trudgill, 2011). This initial pulse of salt migration lasted into the Triassic, and is responsible for lateral thickness variations, truncations, and angular unconformities in the overlying Permian and Triassic strata of the Culter Group, Moenkopi Formation, and Chinle Formation (Doelling et al.,

1988). Subsequently, salt migration diminished and was restricted to localized zones near anticline crests (Doelling et al., 1988). Salt structures were buried by Jurassic sediments and may have been dormant until reactivation from early-Cenozoic (Laramide) shortening, when broad folds following the trends of the Pennsylvanian-Triassic salt anticlines and Pennsylvanian basement faults were formed (Doelling et al., 1988).

Several northwest-striking normal faults developed in the Paleogene during or shortly after the Laramide Orogeny (Olig et al., 1996; Pevear et al., 1997). This period of extensional faulting may be related to relaxation after Laramide compression (McKnight,

1940), reactivation of pre-Paradox faults (Doelling et al., 1988), or to an early phase of epierogenic uplift of the Colorado Plateau (Parker, 1981).

The Moab fault is a 45 km-long, northwest-striking, down-to-the-northeast normal fault that extends from the Tenmile graben to the southern end of Spanish Valley (Olig et al.,

1996). It is subdivided into three sections based on displacement geometry and geomorphic expression, the southernmost of which coincides with the Moab-Spanish

Valley graben (Olig et al., 1996). It’s maximum displacement is ~960 m 5 km northwest of the Colorado River, where it juxtaposes Permian and Upper Jurassic strata (Foxford et

148 al., 1998). The Moab fault is generally interpreted to have formed in Laramide time after

Mesozoic diapirism (McKnight, 1940; Doelling et al., 1988; Olig et al., 1996), through

others suggest that it reactivated older salt structures from the Triassic (Foxford et al.,

1996). Estimates for last displacement on the central strand of the Moab fault range from

1.2-7.5 Ma based on bedrock scarp retreat rates (Olig et al., 1996) to 60 Ma based on Ar-

dating of clay gouge (Solum et al., 2005). Doelling et al. (2002) and Doelling (2004)

inferred that the southern fault section follows the graben axis in Moab and Spanish

valleys where it is concealed beneath Quaternary deposits. This interpretation was based

on well logs and seismic reflection lines in northwest Moab Valley (Cooksley

Geophysics Inc., 1995) and implies that the Moab fault is distinct from the salt-collapse deformation belts on the northeast and southwest graben margins (Olig et al., 1996).

Alternatively, the mapping of Williams (1964) and classification by Black et al. (2004) interpret the Moab fault to bound the southwest margin of the Moab-Spanish Valley graben and to be spatially coincident with any salt-driven deformation. The Moab fault may have a tectonic component, but seismic reflection, aeromagnetic, and gravity data indicate that it soles in the underlying salt anticline and does not penetrate basement rocks (Woodward-Clyde Consultants, 1986; Olig et al., 1996; Trudgill, 2011).

Salt structures throughout the ancestral Paradox Basin are inferred to have been exhumed and reactivated during the late-Cenozoic erosion of the Colorado Plateau

(Doelling et al., 1988). It is thought that dissolution of evaporites by meteoric waters has occurred preferentially along shallowly-buried anticline crests, causing eight of these anticlines to collapse into salt graben (Cater, 1970; Doelling et al., 1988). The Moab salt- cored anticline is one example, which collapsed to form the modern-day Moab-Spanish

149 Valley salt graben. Salt dissolution left behind accumulations of insoluble gypsiferous

residue atop the undissolved salts, which now forms a caprock that is exposed along both

valley walls near the town of Moab, but is hypothesized to be more deeply buried beneath

Spanish Valley (Baars and Doelling, 1987; Doelling et al., 2002).

Some of the only studies to document Quaternary salt deformation in the ancestral

Paradox Basin were conducted around Spanish Valley. Richmond (1962) recognized that

alluvial terraces along Pack Creek converge downstream, and inferred that these units

were deformed by subsidence in Spanish Valley. Harden et al. (1985) extended this

hypothesis, using soil profile analysis to correlate terrace deposits upstream of Spanish

Valley to subsurface gravels and buried soils in the valley floor. They inferred that gravel

deposits observed to converge, thicken, and warp into the subsurface in the middle part of

Spanish Valley dated to the penultimate glaciation, suggesting valley subsidence in the

last 150 ky. More recently, Guerrero et al. (2015) completed a paleoseismic study across

a strand of the Kayenta Heights fault along the northeast graben margin, calculating an

unexpectedly-high mean vertical slip rate of >3 mm/yr with a most recent event post-

dating 2.3 ka. These results suggest that the Moab-Spanish Valley graben is active in the

Holocene, and may be deforming more rapidly, but at less regular intervals, than tectonic

faults elsewhere in the interior western U.S.

QUATERNARY STRATIGRAPHY

Quaternary deposits in the Rill Creek and Kane Springs quadrangles are of

alluvial, colluvial, eolian, and mass-movement origin. Alluvial gravels in the Spanish

Valley subsurface have accumulated to thicknesses of at least 90 m (Lowe et al., 2007).

150 Seven terrace levels of alluvial gravels are present along Mill and Pack creeks upstream of the Moab-Spanish Valley salt graben. On the bedrock plateaus and rims above Spanish

Valley, fine-grain eolian sand derived from the weathering of Mesozoic sandstone is

widespread, and commonly remobilized by overland flow. Talus deposits are common at

the base of escarpments on the margins of Spanish Valley and beneath Wilson and South

mesas to the east.

Alluvial deposits

Qal: Active alluvium (historical and late Holocene). Well to moderately-sorted silt, sand, and gravel comprising the active channel and floodplain of drainages. Deposits are often marked by tamarisk and cottonwood trees. Up to 5 m thick in the canyons of Mill Creek.

Qas: Alluvial sand deposits (Holocene). Moderately-sorted sheetwash deposits not associated with active channels or floodplains. Composed of horizontally-laminated to low-angle cross-bedded sand and silt with minor amounts of clay. Deposits commonly form a mantle up to 3 m thick on alluvial terraces and piedmonts. Alluvial sheetwash sand that is greater than 50% silt and clay is geomorphically distinct and mapped as Qasf.

These deposits are found exclusively in valley bottoms that headwater in Mancos Shale south of Pack Creek.

Qats: Older alluvial sand deposits of Mill and Pack creeks (middle to early Holocene).

Moderately-sorted coarse-grained sand to clay with exposed roots, root casts, and krotovina that forms valley-bottom fill in confined drainages. Unit is mapped in the

canyons of Mill Creek and the entrenched reach of lower Pack Creek between Moab and

Spanish valleys. Deposits form terraces and are cut by arroyos. Deposits locally

151 interfinger with sand to clay slopewash from adjacent valley walls, though most sand is

derived from the mainstem drainage. Exposures along Mill and Pack creeks are 5-15 m

thick.

Qatg1: Older alluvial gravel deposits of Mill and Pack creeks (Holocene to late

Pleistocene, OSL ages of ~6-16 ka). Moderately-sorted, clast-supported, well-rounded, imbricated, pebble-cobble gravel with a matrix of fine to medium-grained sand. Cobble gravel in the eastern map area in the upper reaches of Mill and Pack creeks fines to a pebble gravel downstream in Spanish Valley. Clast lithologies are primarily porphyritic trachyte from the La Sal Mountains and minor locally-derived sandstone and chert.

Deposits contain sparse sand and silt lenses up to 10 m wide. The Qatg1 alluvium comprises much of the modern floor of Spanish Valley, and can be incised 1-5 m by active drainages. In southeast Spanish Valley, the Qatg1 surface forms the Pack Creek alluvial fan. Deposits are 5-12 m thick where exposed in Spanish Valley gravel pits, and thin to 1 m upstream in Mill and Pack creeks. Locally divided into Qatg1y (youngest,

OSL age of ~6 ka), Qatg1i, and Qatg1o (oldest, OSL age of ~16 ka) in upper Pack Creek

according to landscape position of distinct terraces atop the deposit.

Qatg2-Qatg6: Alluvial terrace gravel deposits of Mill and Pack creeks (late to middle

Pleistocene). Moderately-sorted, clast-supported, well-rounded, imbricated, pebble-

cobble gravel to boulder-cobble gravel with a matrix of fine to medium-grained sand.

Clasts are porphyritic trachyte from the La Sal Mountains and lesser locally-derived

sandstone and chert. Qatg2 and Qatg3 contain sparse sand and silt lenses up to 4 m wide,

while sand lenses are rare in Qatg4 and nonexistent in Qatg5 and Qatg6 deposits. Qatg2-

Qatg5 deposits exhibit stage II carbonate rinds in the 1-2 m below terrace treads; Qatg6

152 exhibits stage II-III carbonate development. Qatg2 (OSL ages of ~57-78 ka) treads are

10-25 m above the modern channel of Mill and Pack creeks outside of the Moab-Spanish

Valley graben, and converge with the basin floor along Pack Creek in southeast Spanish

Valley (Fig. 4.3). Qatg2 is not found below Mill Creek’s crossing of the Kayenta Heights

fault, and is hypothesized to be incorporated into the undifferentiated alluvial gravels of

Moab Valley. Deposits form 1-5 m thick strath terraces in the upstream incised reaches of

Mill and Pack Creeks, while along Pack Creek in Spanish Valley they are up to 20 m

thick fill deposits. Qatg3 (OSL ages of ~81 ka) treads are 30-45 m above the modern

channel of Mill and Pack creeks outside of the Moab-Spanish Valley graben and

converge with the basin floor in Spanish Valley similarly to Qatg2. However, Qatg3 deposits extend farther downstream along Pack Creek before projecting into the subsurface in central Spanish Valley (Fig. 4.3). Distribution of Qatg3 deposits along Mill

Creek at the margin of Moab Valley is identical to that of Qatg2. Qatg3 deposits are 2-8 m

thick in the upstream incised reaches of Mill and Pack creeks, but at least 30 m thick fill

deposits in southeast Spanish Valley, where the basal strath is not exposed. Qatg4 (IRSL

age of ~93 ka) treads are 45-70 m above the modern channel in the canyon of South Mill

Creek, which is the only place this unit is found. Deposits form 3-5 m thick strath

terraces. Qatg5 (IRSL age of ~194 ka) deposits are found only as incompletely-preserved

gravel knobs along the canyon walls of South Mill Creek. They are 100-115 m above the

modern channel and 1-4 m thick. Qatg6 (unknown age) deposits are found only along the

northeast graben rim at Johnson’s Up On Top, where they form broad, linear benches and

are partly concealed by a dune field. Treads are 125-145 m above the modern channel of

South Mill Creek and deposits are 10-20 m thick.

153

Figure 4.3. Oblique photo of Pack Creek terraces in southeast Spanish Valley. Terraces T2 (Qatg2) and T3 (Qatg3) converge with the Pack Creek T1 fan (Qatg1) just up-valley from gravel pits. White dashed lines show the tops (treads) of terraces.

Qatg7: Alluvial upland gravel deposits of Johnson’s Ridge and South Mountain (early

Pleistocene, cosmogenic burial ages of ~1.5-1.8 Ma). Moderately to poorly-sorted, clast- supported, poorly stratified, well-rounded, boulder-cobble gravel with a matrix of fine to medium-grained sand. Clast lithologies are primarily porphyritic trachyte from the La Sal

Mountains with sparse locally-derived chert and well-cemented sandstone. Stage III+ pedogenic carbonate development is pervasive, and a stage IV carbonate calcrete occurs in the upper 2 m of the highest, most stable deposits. Clasts on landform treads exhibit a well-developed desert varnish. Deposits are 20-30 m thick and present along the rim of the Moab-Spanish Valley graben at Johnson’s Ridge (east rim) and beneath South

Mountain (south rim). The South Mountain upland gravel tread is 180 m above the modern channel of Pack Creek and the floor of upper Spanish Valley. This deposit is connected to a broad piedmont bajada along the lower flanks of the La Sal Mountains

154 that extends south out of the map area. The Johnson’s Ridge terrace tread is 175 m above

the modern channel of South Mill Creek and 255 m above the floor of central Spanish

Valley. The Johnson’s Ridge upland gravel is a narrow (<300 m wide), north-northwest

trending landform that occupies the divide between Spanish Valley and South Mill

Creek. An equivalent Qatg7 deposit of similar dimensions is located downstream across from the northern extent of Johnson’s Up On Top, and is on the east side of South Mill

Creek.

Qagu: Alluvial gravel deposits of Mill and Pack creeks, undifferentiated (late to middle

Pleistocene, luminescence ages of ~106-137 ka). Same sedimentology and clast lithology

to Qatg2-Qatg5. Deposits form a vertically-continuous gravel package along the eastern

flank of Moab Valley against the Kayenta Heights fault with a terrace is 30 m higher than

the modern valley floor and Qatg1 surface (Fig. 4.4). Imbrication suggests clasts are

derived from Mill Creek. Qagu deposits lack buried soils, and the basal contact is not

visible in the 30 m of exposure. Deposits are concealed by Qatg1 alluvium in Spanish

Valley, and are mapped only on Plate 3.

Qaps: Alluvial piedmont sand deposits (Holocene). Moderately to poorly-sorted coarse- grained sand to clay deposits that form small fans at the mouths of ephemeral piedmont drainages. Deposits occur exclusively along the southwest margin of Spanish Valley, where they grade to the valley floor and are up to 5 m thick.

Qapg1-Qapg5: Alluvial piedmont gravel deposits (Holocene to Pleistocene). Moderately

to poorly-sorted, clast-supported, sub-angular to sub-rounded, boulder-cobble gravel and

pebble-cobble gravel with a matrix of fine to coarse-grained sand. Clast lithologies are

155

Figure 4.4. Undifferentiated alluvium (Qagu) along Mill Creek at the edge of Moab Valley. Qagu deposits are juxtaposed against bedrock along the Kayenta Heights fault zone, and are abruptly truncated on their downstream ends above the T1 (Qatg1) valley floor. Exposed thickness of Qagu shown in photo is 20-30 m.

primarily porphyritic trachyte from the La Sal Mountains and lesser locally derived

sandstone from lower and middle-Jurassic strata, chert from the Morrison Formation, and paraquartzite from the Burro Canyon Formation. Deposits originate from tributaries and valley margins and grade to or prograde onto the corresponding terrace surface (Qatg1-

Qatg5), and are thus inferred to be coeval with, or shortly post-date, terrace ages.

Piedmonts Qapg1 and Qapg2 often form continuous alluvial fans, whereas the older piedmont landforms are often perched above the valley floor and are not always connected to a terrace deposit, especially in the southern half of the map area. Piedmonts

Qapg2-Qapg5 tend to be coarser than Qapg1, and can contain lenses of cobble-boulder gravel.Qapg1 is up to 5 m thick where exposed, and is locally divided into Qapg1y

(youngest) according to landscape position. Qapg2-Qapg5 deposits are up to 20 m thick.

Qafd: Debris fan deposits (Holocene to late Pleistocene). Poorly-sorted, angular to sub-

angular, matrix-supported boulders to sand. Deposits form fans extending from the

156 mouths of steep tributaries along the southwest escarpment of Spanish Valley, and are up

to 4 m thick.

Colluvial deposits

Qc: Colluvium (Holocene to late Pleistocene). Poorly-sorted, clast-supported cobbles and

boulders in a sandy and pebbly matrix that form a thin mantle on slopes below older

Quaternary deposits. Clasts are derived from the reworking of higher terraces, piedmonts,

and talus deposits through raveling and creep, and are thus the same composition as the

source deposit. Colluvial deposits commonly occur along terrace risers, where they

obscure the internal terrace stratigraphy. Deposits are typically less than 4 m thick.

Colluvium is locally divided into Qcy (active/Holocene) and Qco (relict/late-Pleistocene)

based on landscape position and connectivity with the source deposit.

Qca: Mixed colluvial and alluvial deposits (Holocene). Poorly to moderately-sorted,

angular to sub-rounded boulders to sand, but most commonly pebbles, sand, and silt.

Deposits occur at the base of hillslopes, and represent colluvium or talus that has been

reworked by sheetwash and transported downslope. Mixed colluvium and alluvium is

most commonly found at the base of terrace risers along Pack Creek in the southeast map

quadrant, where the deposits slope gently to the valley bottom and are up to 3 m thick.

Eolian deposits

Qes: Eolian sand deposits (Holocene). Well-sorted, quartz-rich, fine-grained sand and silt that forms coppice dunes and sand sheets in open valley bottoms and upland surfaces, and falling dunes on the lee side of bedrock ledges. Deposits commonly mantle or are in pockets downwind of Navajo Sandstone outcrops. Dunes are rarely exposed in cross-

157 section, but show approximately 1 m tangential cross beds in gully walls. Where

vegetated, dunes are massive. Sand sheets along valley bottoms and in the lee of bedrock

protrusions are generally less than 5 m thick, though a few deposits are locally 20 m thick

in bedrock embayments in southeast Spanish Valley.

Qeas: Mixed eolian and alluvial sand deposits (Holocene). Well to moderately-sorted

sheets of fine to medium-grained sand and silt deposited by wind and reworked by

sheetwash. Deposits commonly mantle alluvial terrace and piedmont deposits, form a

thin cover on bedrock plateaus above the rim of Spanish Valley, fill large embayments

along the foot of the graben escarpments, and form gently sloping surfaces between

bedrock canyon walls and active washes in the Mill Creek drainage. Deposits are rarely

exposed in profile, but may display tangential cross beds to horizontal laminations

depending on the intensity of alluvial reworking. Qeas is distinguished from Qas by the

presence of eolian cross beds and greater topographic roughness, and differentiated from

Qes by the absence of dune topography. Thickness varies from 1-2 m atop terraces and

piedmonts to 12 m at the foot of escarpments.

Mass-movement deposits

Qmt: Talus (Holocene to late Pleistocene). Poorly-sorted, angular, clast-supported, cobbles to 5 m-diameter boulders in a sandy matrix. Clasts are derived from gravity-

driven mass movements and are predominantly locally-sourced sandstone. Talus cones

and aprons most commonly mantle slopes at the base of cliffs along the southwest

Spanish Valley escarpment and below Wilson and South mesas. Deposits are up to 8 m

thick, and can grade downslope to colluvium and mixed colluvium/alluvium.

158 Qms: Landslide deposits (late to early Pleistocene). Unconsolidated to partially-intact

slump blocks of the Brushy Basin member of the Morrison Formation and the overlying

Burro Canyon Formation. Intact bedrock blocks exhibit inconsistent bedding orientations,

and can be displaced several hundred meters down-slope. Landslides are found only in

the southern map area, where there are extensive Brushy Basin outcrops. The largest

landslide deposit is in extreme southeast Spanish Valley north of Pack Creek and below

2 an adjacent Qapg4 piedmont. It is 0.4 km in area with a thickness likely greater than 40 m.

Qmbx: Fissure-fill/collapse breccia deposits (Pleistocene). Unconsolidated, poorly sorted, angular pebble to boulder deposits that have been displaced from stratigraphic levels 50 to 700 m (Weir et al., 1994), averaging 200 m. In most cases, the source lithology has been eroded from the local setting. Deposits form 30-250 m wide conical

hills or linear ridges aligned with faults, and are only found on the northeast margin of

Spanish Valley. Deposits lie unconformably atop in-situ bedrock, which is most

commonly Navajo Sandstone (Fig. 4.5). The basal bedrock contact is often, but not

always, exposed, and visible thicknesses range from 3 to 40 m. Deposits are subdivided

based on the youngest clast lithology present. Qmbxjm contains clasts from stratigraphic

units up to the Brushy Basin member of the Morrison, and is distinguishable by the

presence of green claystone pebbles and granules, chert pebbles to boulders, and coarse

fluvial sandstone boulders. Qmbxkm contains clasts from stratigraphic units up to the

Mancos Shale, and is found only below a Burro Canyon Formation ridge in the

monocline north of Pack Creek in the southeast map quadrant. Qmbxqg contains rounded

159 porphyritic trachyte alluvial gravels from the La Sal Mountains interpreted to be

displaced from higher terrace and piedmont deposits.

Figure 4.5. Fissure-fill/collapse breccia along the northeast graben margin below Johnson’s Ridge. Deposit shown here contains unconsolidated clasts from the Morrison Formation and is mapped as Qmbxjm. Note spatial association with mapped fault and basal unconformity with bedrock.

Disturbed deposits

Qh: Anthropogenic excavations, fill, and disturbed deposits (historical). Disturbed deposits include gravel pits, landfills, reservoirs, and earthen dam retention basins, which are found primarily in central and southern Spanish Valley.

160 STRUCTURE/NEOTECTONICS

The dominant structural features in the map area are found along the flanks of the

Spanish Valley graben. Graben-bounding fault systems with a normal sense of displacement vary significantly in their geometry between the two margins. The Moab fault zone occupies the southwest graben margin and the Kayenta Heights fault zone

(Doelling et al., 2002) comprises the northeast margin. The structural geology outside of the graben is dominated by strata with a 5-10° southwest regional dip, though there is local variation near the graben rims. Around South Mill Creek and Johnson’s Up On Top, strata dip northeast at 3-17° in a 1-2 km belt extending northeast from the graben.

Bedrock within 5 km of both graben margins is marked by 1-5 m-spaced joints, which

include a primary set that is north-to-northwest-trending and a secondary set that is northeast-to-north trending.

Moab fault zone

The southern section of the Moab fault zone (MFZ) forms the southwest margin of the Moab-Spanish Valley graben (Black et al., 2004; U.S. Geological Survey and Utah

Geological Survey, 2006). The deformation zone is typified by a series of master and subsidiary faults that accommodate strata of the Glen Canyon Group rolling over into the floor of the graben (Plate 2). It includes both synthetic and antithetic faults and locally contains non-pervasive cataclastic deformation bands near master faults. Displacement of exposed units across the fault zone is 250-300 m, though total displacement including concealed fault strands in the floor of the graben is estimated at ~475 m (Plate 2). Master faults are the farthest outboard (southwest) from the graben and form the prominent

161 skyline escarpment southwest of Spanish Valley. Based on offset marker beds they accommodate 200-250 m of displacement. Master faults are broken into sub-segments up to 4 km long with an average length of 2.6 km. Sub-segment boundaries mark locations where faults step inboard or outboard, or where subsidiary faults join master faults. There is a prominent boundary in the map area 1 km south of the Grand/San Juan county line where the master fault to the northwest is 100 m outboard of the master fault to the southeast. To the northwest of this segment boundary an antithetic fault bounds a narrow graben below the master fault, and a second, parallel master fault 300 m inboard of the main escarpment extends to the northwest off the quadrangle.

Subsidiary faults closer to the graben axis are shorter than master faults, averaging 0.7 km in length. Based on offset marker beds in the Kayenta Formation these faults accommodate modest offsets of less than 50 m, display synthetic (down to the northeast) and antithetic (down to the southwest) senses of slip, and separate blocks with dips that progressively steepen towards the graben (Fig. 4.6). Strata in the footwalls of the master faults dip 6-9° south and southeast, while bedrock hogbacks of the master fault hanging walls dip northeast at 25-40° (Plate 1). Quaternary alluvial and eolian deposits onlap the hogbacks at the topographic floor of the graben, and extend into erosional embayments along the valley margin. It is probable that there are additional subsidiary faults closer to the graben axis along the MFZ that are concealed beneath Quaternary alluvium (Plate 2). These inferred faults are required to explain the small outcrops of

Cretaceous Burro Canyon Formation west and south of Ken’s Lake (Plate 1).

The MFZ is less than 400 m wide near its southeast termination, and widens to

1.2 km at the western edge of the map area (Plate 1). The high width along the Moab-

162

Figure 4.6. Master and subsidiary faults in the Moab fault zone. Note the change in dip of the marker bed across the master fault. Photo was taken near the southeast terminus of the MFZ where fault slip lessens and transitions into an unbroken asymmetric anticline (off the photo to the left). Master fault here displays modest throw (~20 m) compared to farther northwest along the fault zone.

Spanish Valley graben distinguishes this southern section of the Moab fault from the central and northern sections, which exhibit a ~10 m wide deformation zone in exposures

north of Moab (Olig et al., 1996; Foxford et al., 1998). Slip on the MFZ decreases to the

southeast, where faults transition to an asymmetric anticline that drapes downward into

uppermost Spanish Valley (Plate 1). The MFZ continues 9.5 km off the map to the

Colorado River in the northwest direction, where caprock of the Paradox Formation is

locally exposed along the base of the escarpment as a fault block (Doelling et al., 2002).

163 Several alternate hypotheses may account for the geometry of the Moab fault zone

in Spanish Valley. They fall into end-member classes of salt-related lateral spreading and

purely structural folding and faulting. A combination of these processes is most likely to

explain the mapped geometry of the MFZ.

First, the progressive steepening of bedding towards the graben axis may record a

spreading and tilting phenomenon, where mapped faults are failure planes that allow

mass movement blocks to slide and fall into the subsiding basin. This process has been

documented in the Needles district of Canyonlands and is hypothesized to result from

either lateral salt flow towards areas of lower confining pressure in Cataract Canyon

(Baker, 1933) or basal sliding along a tilted salt plane (Huntoon, 1982). The later hypothesis requires a basal salt plane angled down towards the graben axis, which is not supported by seismic, gravity, or well data for Moab and Spanish valleys (e.g. Case and

Joesting, 1972; Trudgill, 2011). The lateral spreading process in the Needles fault zone has produced a deranged drainage pattern and unique fissure-fill deposits termed

“swallow holes” (Geiger, 2014). The complete absence of fissure-fills in the MFZ argues against the spreading and tilting model, as extension would be expected to produce open fractures that collect overburden. Further, if opening fissures had in fact collected sediment, at least some of these deposits should be preserved in the landscape – either still in-situ or as topographically-inverted hills. That such fissure-fill deposits, composed of soft shale from the Brushy Basin Member of the Morrison Formation, are documented in the KHFZ on the other side of the graben indicates that the lack of fissure-fills in the

MFZ is not simply a product of low preservation potential.

164 Alternatively, purely structural processes may explain some components of the

MFZ geometry in Spanish Valley. The MFZ may be taken to represent a partially- exhumed fault propagation fold where strata that drape into the graben are cut by subsidiary faults that connect at depth to a master fault (e.g. Sharp et al., 2000). The asymmetric anticline that marks the southeast end of the fault zone appears to be cored by faults that have not yet breached the surface, which fits the conceptual model for a fault propagation fold (Fig. 4.6). This interpretation would imply that increased slip to the northwest along the graben margin has in fact propagated to the surface as brittle deformation, resulting in the mapped pattern of multiple faults cutting strata that roll over into the graben. However, the fault propagation fold hypothesis does not completely match observations of deformation along the MFZ. Fault propagation folds in extensional settings are typically portrayed as having subsidiary faults that converge with a master fault at depth (Sharp et al., 2000), yet these steeply-dipping faults may not all connect at

depth (Plate 2) as inferred by Doelling et al. (2002) in Moab Valley.

The preferred interpretation for the mapped geometry of the Moab fault zone is a

combination of salt tectonic and more purely structural processes that includes

components of the tilt-block and fault propagation fold end-members. It seems most

likely that faults in the MFZ are a product of dissolution-induced subsidence rather than

extension. As subsurface dissolution removes salt in the Paradox Formation, the

overlying strata are displaced downward along discrete, near-vertical failure planes. The

post-Paradox strata are left behind as foundered blocks amongst the insoluble caprock

and are separated by synthetic and antithetic faults in the floor of the graben (Plate 2).

This may explain the isolated bedrock outcrops near Ken’s Lake (Plate 1) and well data

165 showing that alluvium in the floor of Spanish Valley is underlain by Mesozoic bedrock

(Lowe et al., 2007). It is interpreted that the graben-margin faults accommodating subsidence are shallowly cored in the Paradox Formation (Doelling et al., 2002;

Gutiérrez, 2004), which well data indicates is no deeper than 600 m beneath the flank of

Spanish Valley (UDOGM, 1973). Near the southeast limit of the graben, salt-dissolution

subsidence has not been great enough to cause brittle faulting at the surface and instead

strata are folded down towards the graben floor in a hinge zone (Plate 1). It is possible

that faults with decreasing slip exist beneath the surface of this fold and die out with the

southeast limit of salt dissolution. In summary, current faults in the wide, heterogeneous

southern section of the Moab fault zone are interpreted as gravity-driven, shallowly-

rooted, and non-tectonic, which is the same as inferred by many prior workers (Cater,

1970; Doelling et al., 1988; Olig et al., 1996; Doelling et al., 2002; Gutiérrez, 2004;

Trudgill, 2011; Guerrero et al., 2015).

Kayenta Heights fault zone

The Kayenta Heights fault zone (KHFZ) is characterized by linear, up to 80°

dipping, down-to-the-southwest normal faults. The footwall of the fault zone is Navajo

Sandstone, and the downthrown block is mostly buried beneath Quaternary alluvial

deposits of Spanish Valley (Plate 1). The fault zone is 1.5-2 km wide in northwest and

central Spanish Valley and narrows to 0.5 km as slip lessens to the southeast. The

southeast extent of the fault zone is captured in the map area by the transition from

northwest-trending faults to a west-northwest trending asymmetric anticline dipping

steeply to the south-southwest along upper Pack Creek (Plate 1). The KHFZ continues 9

166 km northwest to the Colorado River off the western boundary of the map (Doelling et al.,

2002).

Faults of the KHFZ parallel pervasive northwest-trending joints in the Navajo

Sandstone that are spaced at 1-5 m along the graben margin. Unlike the MFZ, cataclastic

deformation bands are ubiquitous throughout the KHFZ (Fig. 4.7A) and their greatest

density is not always adjacent to fault strands. Bedrock along faults and fractures is

commonly brecciated and bleached (Fig. 4.7B), suggesting strain localization and fluid

alteration. Because cataclastic bands occur throughout the deformation zone and not just

adjacent to mapped faults, it is possible they are inherited from an earlier phase of

deformation. Their presence along the KHFZ and near absence along the MFZ through

Spanish Valley may also record differing deformation styles between the two graben

margins. However, it is equally as likely that this discrepancy is caused by lithologic

variations between the Navajo Sandstone along the northeast margin and the Kayenta

Formation and Wingate Sandstone along the southwest margin (Plate 1), as mechanical

differences between rock units can affect the density of deformation bands (Fossen et al.,

2007).

The main footwall block of the KHFZ is comprised of 2-3 parallel faults tens to

hundreds of meters apart. Total displacement across the exposed width of the fault zone

is ~200 m, though it is likely that there are additional faults in the hangingwall block

concealed beneath Quaternary alluvium (Plate 2). Individual fault strands are up to 3 km

long with an average length of only 0.7 km, and accommodate 10 to >120 m of offset.

Sense of motion on all faults is down-to-the-southwest, and there are no antithetic faults.

The KHFZ includes several segment boundaries where the locus of deformation steps

167

Figure 4.7. Kayenta Heights fault zone deformation textures. (A) Fault breccia and cataclastic deformation bands in the Navajo Sandstone. Location: 38.47946°, - 109.40881°. (B) Bleaching and brecciation in the Moab Member of the Curtis Formation. Location: 38.46011°, -109.39178°. Scale is 15 cm long in both images.

inboard (towards the graben) or outboard (away from the graben) or where fault planes change strike. These are located east of Ken’s Lake, near the Johnson’s Up On Top access road, at the northern Spanish Valley water tank (marked “WT” on the Rill Creek quadrangle), and north of the Moab Golf Club where the KHFZ steps 1 km to the west in the down-valley direction (Plate 1). The northern Spanish Valley water tank segment boundary exhibits a well-developed fault relay ramp. Where there are parallel faults along the KHFZ, the greatest displacement is along the fault closest to the graben (Plate

2), suggesting that concealed faults farther inboard may accommodate even more displacement. In central Spanish Valley this inboard-most visible fault is marked by a 1-2 m wide zone of fault breccia, and follows the straight, vertical cliff that emerges abruptly

168 from the Quaternary deposits of the basin floor. The displacement along this escarpment must be greater than the 120 m of topographic relief, as the bedrock in the hangingwall is buried beneath eolian and alluvial deposits. In the northwest map quadrant a strand of the

KHFZ is confined within the Navajo Sandstone in both the hangingwall and the footwall but inferred to have only 30-40 m of throw (Guerrero et al., 2015). Here, an inferred fault

200-400 m inboard (to the southwest) juxtaposes the Salt Wash and Brushy Basin members of the Morrison against Navajo Sandstone in the footwall (Plate 1), and thus produces a much greater displacement of at least 150 m.

Conical hills and linear ridges of fissure-fill breccia (Qmbx) are common in the

KHFZ, but are found nowhere else in the map area – including the MFZ. These deposits occur along or adjacent to faults, and lie unconformably on in-situ bedrock (Fig. 4.5).

Because the basal bedrock contact is almost always visible, these deposits were not formed by sinkhole collapse in the Paradox Formation as suggested by Weir et al. (1994).

Instead, these features represent topographic inversion of material that fell into fissures within the Navajo Sandstone during extension. They are most commonly composed of clasts from the Morrison Formation (Qmbxjm) displaced 150-250 m down-section, and overlie well-jointed but otherwise underformed Navajo Sandstone. Fissure-fill breccias are interpreted as relict analogs to the “swallow holes” of the Needles fault zone (Geiger,

2014).

Morrison Formation clasts in fissure-fill deposits may have come from either the footwall (northeast) or hangingwall (southwest) side of the KHFZ. If these clasts are derived from the footwall, fissure-fill breccias must have formed before the cliffs of the upper Jurassic section retreated to the northeast, as the Morrison Formation is no longer

169 present along the northeast footwall rim of Spanish Valley. This would have occurred

before the deposition of the Qatg7 Johnson’s Ridge gravel, which is inset below

exposures of the Morrison (Plate 1). Alternatively, the Morrison Formation clasts in

Qmbx deposits may have come from the hangingwall of the KHFZ during graben

subsidence. Morrison is exposed in the hangingwall near the eastern boundary of the

quadrangle and is present in the asymmetric anticline at the southeast end of the KHFZ

(Plate 1). The remainder of the hangingwall along the KHFZ is concealed beneath

Quaternary alluvium and it is unknown if Morrison is buried along the flanks of Spanish

Valley. The hangingwall-origin hypothesis implies that Qmbx deposits are coeval with

graben subsidence, which is supported by their close spatial association with faults of the

KHFZ that accommodate subsidence (Plate 1). This hypothesis offers a simpler

explanation for fissure-fill deposits than a footwall origin, though mapping presented here

cannot definitively confirm the origin of these unique features. Regardless, the period of

fissure-fill deposition may have been distinct from the modern subsidence of the Moab-

Spanish Valley graben. While there are joints and fractures along the KHFZ today, there

are no modern equivalents to the >100 m wide fissures, and no evidence for active

deposition of overburden into these fissures, like there must have been during the

formation of Qmbx deposits.

Implications for graben evolution and subsurface geometry

The differences in geometry between the Kayenta Heights and Moab fault zones offer insights into the evolution of the Moab-Spanish Valley salt graben. Modern deformation along both graben margins appears to be controlled by vertical gravitational

170 collapse from the subsurface removal of dissolved salts. Mapped fault strands are high-

angled (up to 80° dipping), inferred to be shallowly-rooted in the underlying Paradox salt

wall (Plate 2), and lack modern fissure-fill analogs that would imply an extensional component. However, previous workers have inferred that these modern mapped deformation belts related to Quaternary salt-dissolution subsidence are distinct from the pre-collapse deformation and geometry along the crest of the Moab salt anticline (e.g.

Olig et al., 1996; Doelling et al., 2002). Mapping and field observations presented here demonstrate that there may have been differences in deformation style between the Moab fault zone and what is now the Kayenta Heights fault zone prior to the ongoing period of salt-dissolution subsidence observed today. There is 350 m of structural relief at the base of the Navajo Sandstone between the southwest (higher) and northeast (lower) graben margins (Plate 2). This relief is inferred to represent the pre-collapse offset along the southern section of the Moab fault, of which the geometry and deformation style is uncertain. Furthermore, the fissure-fill breccias of the KHFZ imply a period of extension and fracture-opening that is not observed along the mapped Moab fault zone (Plate 1), which may instead reflect pre-collapse or early-collapse deformation in the hangingwall of the MFZ. Collectively, these observations support but cannot confirm the interpretation of Olig et al. (1996) that Quaternary salt-dissolution subsidence accommodated along the MFZ and KHFZ is distinct in time and structural style from pre- collapse deformation along the Moab fault zone.

The specific relationship between modern graben subsidence and pre-collapse deformation on the Moab fault zone remains poorly-understood. Prior to salt-dissolution collapse, the southern section of the Moab fault zone may have behaved much like the

171 central section northwest of Moab, which is characterized by a narrow (<10 m)

deformation zone and the presence of cataclastic deformation bands (Olig et al., 1996;

Foxford et al., 1998; Fossen et al., 2005). Because it formed along, and possibly

reactivated structures related to, the initial Moab salt anticline (Doelling et al., 1988), it is

inevitable that the southern section of the Moab fault zone acted as a salt conduit and

focus of collapse during the early period of salt dissolution. It is inferred that with the

onset of Quaternary salt-dissolution subsidence, deformation along the southern section

of the Moab fault changed to a wider (>1 km), more heterogenous fault zone composed

of short strands that are shallowly-cored in the Paradox Formation while the central

segment (where the adjacent salt anticline was not breached) remained relatively unchanged (Olig et al., 1996). It is debated where the pre-dissolution trace of the Moab

fault’s southern segment was located in relation to the Moab salt anticline, and whether

evidence of this trace is still preserved. The trace is variously mapped as following the

modern-day axis of Moab and Spanish valleys beneath the alluvial cover (Doelling et al.,

2002; Doelling, 2004), coinciding with the northeast extent of the salt wall and following

the northeast graben margin (Banbury, 2005), or following the modern southwest

escarpment of Spanish Valley (Baker, 1933; Williams, 1964; Black et al., 2004).

2D seismic data across Moab Valley and northern Spanish Valley show a

symmetrical salt wall and ~400 m of structural relief between Mesozoic strata on either

side of the graben, but do not resolve the geometry of strata known to overlie caprock in

Spanish Valley (Lowe et al., 2007; Trudgill, 2011 Fig. 7). Gravity data may provide

additional information as to whether a pre-collapse trace of the southern segment of the

Moab fault juxtaposes salt against Permian through upper Jurassic strata beneath the

172 alluvium of Spanish Valley (Doelling et al., 2002; Doelling, 2004). An abrupt gradient from a lower gravity anomaly to the southwest (in the footwall of the pre-dissolution

Moab fault) to a higher gravity anomaly to the northeast (in the hangingwall where a thick package of denser sedimentary rocks overlie evaporites) would be the expected signal for this hypothesis, while a more symmetrical gravity low would argue against a major fault in the subsurface. A gravity anomaly cross section from Case and Joesting

(1972) shows a symmetrical low over Moab Valley superimposed on a regional field that decreases to the northeast. The gravity survey of Hildebrand and Kucks (1983) lacks data points in and around Spanish Valley, but interpolated contours show a symmetrical gravity low. Data from the World Gravity Map (Bonvalot et al., 2012) have a 2 km resolution that is too coarse to distinguish gravity differences in the bottom of Spanish

Valley.

Based on the balance of this limited subsurface evidence, the cross section in

Plate 2 was drawn with a roughly-symmetrical salt wall and without a singular, cross- cutting, pre-dissolution, Moab fault plane. Instead, the structural relief between the southwest and northeast graben margins is interpreted as being accommodated across a system of concealed faults that separate horst and graben blocks that are foundered within the Paradox caprock (Plate 2). A faulted fold near the graben axis is inferred based on the nearest mid-valley bedrock exposures 1.8 km southeast of the cross section line, where bedding dips ~40 west-southwest (Plate 1). It is possible that a fault along the graben axis associated with the southern MFZ provides a structural control to the south and southwest-dipping Cretaceous strata north of Pack Creek as a possible rollover anticline

(Plate 1). A similar rollover geometry is documented in the hangingwall of the Moab

173 fault north of Moab Valley (Doelling et al., 2002), though the presence of this pattern in

southernmost Spanish Valley is unconfirmed by subsurface evidence.

GEOLOGIC HAZARDS

Faulting and seismicity

Active subsidence of the Moab-Spanish Valley salt graben is accommodated along structures that cannot be considered true tectonic faults. Interpreted as shallow- rooted features that extend only hundreds of meters into the Paradox Formation, these faults likely do not accumulate and release strain in large seismic events (Olig et al.,

1996; Hylland and Mulvey, 2003; Gutiérrez, 2004) and are thus classified as Class B structures in the U.S. Quaternary Faults and Folds Database (Black et al., 2004; U.S.

Geological Survey and Utah Geological Survey, 2006). This interpretation is consistent with the paleoseismic study of Guerrero et al. (2015), who found that deformation along the northeast margin of Moab Valley is both episodic and anomalously rapid (greater than 3 mm/yr), yet has a much shorter average recurrence interval ( ~300 years) that is more stochastic than tectonic faults. This suggests that even though rates of salt deformation may be as rapid as along tectonic structures, faults along the KHFZ and

MFZ likely do not pose the same ground-shaking hazard.

The paleoseismic analysis of Guerrero et al. (2015) and the mapping and geochronology along Mill Creek at the entrance to Moab Valley from this study (Plate 3;

Chapter 2) indicate that there has been displacement across the Kayenta Heights fault zone in the late Pleistocene and Holocene. Using Qatg4 age-equivalent terrace gravels and undifferentiated alluvial deposits on either side of the KHFZ, the slip rate is

174 constrained to 439 ± 49 m/my (0.439 ± 0.049 mm/yr) since the middle to late Pleistocene

(Chapter 2, Fig. 2.14). This geochronological data for displaced late-Pleistocene deposits

is a development from the previous geologic hazards study of Moab Valley, which cited

no evidence for displacement along the graben-bounding fault zones over this time period

(Hylland and Mulvey, 2003). If the southern section of the Moab fault zone along the

southwest graben margin is related to the KHFZ, the lack of displaced Quaternary units

along the MFZ may reflect a period of inactivity but does not rule out future slip events.

This new evidence for late-Pleistocene activity along the KHFZ suggests that there is a

potential surface-rupture hazard associated with graben-margin salt-dissolution faults for

overlying structures and lifelines, though the conclusions of Hylland and Mulvey (2003)

inferring a low ground-shaking hazard from these faults remain valid.

The primary tectonic structure in the region is the Moab fault, which pre-dates salt-dissolution subsidence of the Moab-Spanish Valley graben and extends 35 km northwest beyond the map area (Olig et al., 1996). Though its early history includes a tectonic phase (McKnight, 1940; Doelling et al., 1988), it is likely that that much of its deformation has involved salt because of its position adjacent to the Moab anticline, its root in the Paradox Formation, and the regional Mesozoic and Cenozoic salt mobility documented by stratigraphic unconformities (Woodward-Clyde Consultants, 1986;

Trudgill, 2011). At a minimum, it is generally agreed that the current manifestation of the

Moab fault is modified by salt tectonism (e.g. Olig et al., 1996; Doelling et al., 2002;

Banbury, 2005; Trudgill, 2011). The mobility and lateral flow of salt (Trudgill, 2011) has produced different structural features along strike of the Moab fault (Olig et al., 1996) and makes it challenging to infer the structural and tectonic history of the fault zone.

175 Studies of the Moab fault ~20 km northwest of the map area conclude that the last period

of major faulting was in the Paleogene (Pevear et al., 1997; Garden et al., 2001; Solum et al., 2005). However, it is possible that the southern section of the Moab fault through

Spanish Valley has had a different history, as its fault rocks have not been dated. There is no microseismicity recognized along the Moab fault (Wong and Simon, 1981; Wong et al., 1996) and historic earthquakes greater than magnitude 4 are very rare in the Paradox

Basin (Wong and Humphrey, 1989). The nearest recent notable earthquakes to Spanish

Valley have occurred in the Paradox Valley ~55 km southeast. Here, five earthquakes greater than magnitude 3 have occurred since 2000, with the largest being a magnitude

4.3 (“USGS Earthquake Hazards Program”). In summary, mapping presented here agrees with prior studies inferring that the southern section of the Moab fault zone is non- tectonic and poses a low earthquake hazard for Moab and Spanish valleys (Olig et al.,

1996; Wong et al., 1996; Hylland and Mulvey, 2003).

Rock fall and other mass movement

Rock fall is a more significant geologic hazard associated with salt-dissolution faulting and collapse along the margins of Spanish Valley. Rock fall is active in other parts of the map area, such as the canyons of Mill Creek and the cliffs below Wilson and

South mesas, but poses the greatest hazard to humans in the developed portions of

Spanish Valley. As the valley floor continues to subside, cliffs forming the northeast and southwest escarpments of Spanish Valley will continue to retreat and increase in relief.

Rock fall is most probable along the southwest graben margin, where there is 500 m of topographic relief and the MFZ footwall is pervasively jointed parallel to the escarpment

176 free-face (Hylland and Mulvey, 2003). Extensive talus deposits of boulders from Glen

Canyon Group sandstone are mapped along this valley margin (Plate 1). These deposits

are more widespread to the northwest, where development along U.S. Highway 191 is

encroaching from Moab, and several talus cones overlie alluvial and eolian deposits

along the valley floor. Rock fall in the southwest wall of Moab and Spanish valleys is an

active hazard, as confirmed by a 2014 rock fall from the Wingate Sandstone that blocked

the entrance to a railroad tunnel just north of the Colorado River with an estimated 4500

m3 of material (Office of Environmental Management, 2015). The lesser topographic relief and scarcity of talus deposits suggests there is a lower likelihood of rock fall along the northeast valley margin, though active subdivision development below the linear

Navajo Sandstone escarpment between Ken’s Lake and the Moab Golf Club increases the probability of adverse impacts. Debris flows occur along the southwest wall of Spanish

Valley as evidenced by unsorted, boulder-rich fans (Plate 1). These deposits extend less than 200 m from the base of the escarpment, and thus only pose a hazard to immediately- adjacent structures, of which there are none in the map area. This analysis agrees with the debris flow and rock fall hazard map of Hylland and Mulvey (2003), which shows that the greatest hazard is along and at the base of the graben escarpments. To mitigate the societal risks of rock fall and other mass movement events, development in Spanish

Valley should occur well away from the valley walls, especially in areas where there are mapped talus deposits.

177 Ground subsidence

The imperceptible, large-scale subsidence occurring over the broad area of the

Moab-Spanish Valley graben does not itself pose a notable geologic hazard. Subsidence

throughout the late Pleistocene and Holocene in Spanish Valley has averaged

approximately 0.5 mm/yr (Chapter 2), which is likely not enough to create foundation

problems for structures in the alluvial valley bottom. Smaller-scale, more rapid

dissolution events are only a concern where evaporites are near the surface. There are

exposures of Paradox Formation caprock on the flanks of Moab Valley (Doelling et al.,

2002), where dissolution produced a sinkhole collapse in 2015 (Grant Willis, personal

communication). There are no exposures of the Paradox Formation in Spanish Valley or

anywhere else in the Rill Creek/Kane Springs map area, though the unit may be within

250 m of the surface on the northeast edge of the MFZ (Plate 2). This is likely deep

enough to not pose an acute dissolution hazard to structures above. The hazard maps of

Hylland and Mulvey (2003) agree with this assessment, as they map the acute dissolution hazard from expansive soil and rock as ending west of the Rill Creek quadrangle boundary (Plate 1), whereas the broader-scale zone of valley- subsidence (which is a low

hazard) extends throughout the alluvial floor of Moab and Spanish valleys (Plate 4).

Other soil and dissolution hazards recognized in Moab Valley that relate to clay-rich

lithology in the Chinle and Moenkopi formations (Doelling et al., 2002; Hylland and

Mulvey, 2003) are not present in Spanish Valley, as Triassic strata are not exposed at the surface in the map area. Bentonite clay is present in the Brushy Basin Member of the

Morrison Formation, which outcrops near the southeast end of Spanish Valley (Plate 1).

This clay is susceptible to shrink/swell cycles when wetted and may destabilize structures

178 built on it, of which there are only a few west of Brumley Creek. To minimize the human

hazards of salt-dissolution subsidence and expandable soils, structures should not be built

directly on Paradox Formation caprock or Mesozoic clay-rich units.

REFERENCES

Baars, D.L., and Doelling, H.H., 1987, Moab salt-intruded anticline, east-central Utah: Geological Society of America Centennial Field Guide, Rocky Mountain Section, p. 275–280.

Baker, A.A., 1933, Geology and oil possibilities of the Moab district, Grand and San Juan counties, Utah: U.S. Geological Survey Bulletin 841, 95 p., https://pubs.usgs.gov/bul/0841/report.pdf.

Banbury, N.J., 2005, The role of salt mobility in the development of supra-salt sedimentary depocenters and structural styles [Ph.D. thesis]: University of Edinburgh, 408 p.

Bonvalot, S., Balmino, G., Briais, A., Kuhn, M., Peyrefitte, A., Vales, N., Biancale, R., Gabalda, G., Reinquin, F., and Sarrailh, M., 2012, World Gravity Map: Commission for the Geological Map of the World.

Case, J.E., and Joesting, H.R., 1972, Regional geophysical investigations in the central Colorado Plateau: U.S. Geological Survey Professional Paper 736, 31 p., http://pubs.er.usgs.gov/publication/pp736 (accessed February 2018).

Cater, F.W., 1970, Geology of the salt anticline region in southwestern Colorado: U.S. Geological Survey Professional Paper 637, 80 p., https://pubs.usgs.gov/pp/0637/report.pdf.

Cooksley Geophysics Inc., 1995, Reflection seismic survey of Atlas Minerals tailing site three miles northwest of Moab, Utah: unpublished consultants report:, 12 p.

179 Doelling, H.H., 2004, Geologic Map of the La Sal 30’ x 60’ Quadrangle, San Juan, Wayne, and Garfield Counties, Utah, and Montrose and San Miguel Counties, Colorado: Utah Geological Survey.

Doelling, H.H., 2001, Geologic map of the Moab and eastern part of the San Rafael Desert 30’ x 60’ quadrangles, Grand and Emery Counties, Utah, and Mesa County, Colorado: Utah Geological Survey.

Doelling, H.H., Oviatt, C.G., and Huntoon, P.W., 1988, Salt deformation in the Paradox region: Utah Geological Survey, v. Bulletin 122.

Doelling, H.H., Ross, M.L., and Mulvey, W.E., 2002, Geologic Map of the Moab 7.5’ Quadrangle, Grand County, Utah: Utah Geological Survey.

Fossen, H., Johansen, T.E.S., Hesthammer, J., and Rotevatn, A., 2005, Fault interaction in porous sandstone and implications for reservoir management; examples from southern Utah: AAPG Bulletin, v. 89, p. 1593–1606, doi: 10.1306/07290505041.

Fossen, H., Schultz, R.A., Shipton, Z.K., and Mair, K., 2007, Deformation bands in sandstone: a review: Journal of the Geological Society, v. 164, p. 755–769, doi: 10.1144/0016-76492006-036.

Foxford, K.A., Garden, I.R., Guscott, S.C., Burley, S.D., Lewis, J.J.M., Walsh, J.J., and Watterson, J., 1996, The Field Geology of the Moab Fault: , p. 265–284.

Foxford, K.A., Walsh, J.J., Watterson, J., Garden, I.R., Guscott, S.C., and Burley, S.D., 1998, Structure and content of the Moab Fault Zone, Utah, USA, and its implications for fault seal prediction: Geological Society, London, Special Publications, v. 147, p. 87–103, doi: 10.1144/GSL.SP.1998.147.01.06.

Garden, I.R., Guscott, S.C., Burley, S.D., Foxford, K.A., Walsh, J.J., and Marshall, J., 2001, An exhumed palaeo-hydrocarbon migration fairway in a faulted carrier system, Entrada Sandstone of SE Utah, USA: Geofluids, v. 1, p. 195–213, doi: 10.1046/j.1468-8123.2001.00018.x.

180 Hackman, R.J., 1956, Photogeologic map of the Mount Peale-4 Quadrangle, San Juan County, Utah: U.S. Geological Survey.

Harden, D.R., Biggar, N.E., and Gillam, M.L., 1985, Quaternary deposits and soils in and around Spanish Valley, Utah: GSA Special Papers, v. 203, p. 43–64, doi: https://dx.doi.org/10.1130/SPE203.

Hildebrand, T.G., and Kucks, R.P., 1983, Regional magnetic and gravity features of the Gibson Dome area and surrounding region, Paradox Basin, Utah: A preliminary report.: U.S. Geological Survey Open-File Report 83–359.

Hylland, M.D., and Mulvey, W.E., 2003, Geologic hazards of Moab-Spanish Valley, Grand County, Utah: Utah Geological Survey Special Study 107, 34 p., http://digitallibrary.utah.gov/awweb/awarchive?item=42825 (accessed February 2018).

McKnight, E.T., 1940, Geology of area between Green and Colorado rivers, Grand and San Juan counties, Utah: U.S. Geological Survey Bulletin 908, 147 p., https://pubs.er.usgs.gov/publication/b908.

Office of Environmental Management, 2015, Moab Resumes Rail Shipments After Rockslide: Energy.gov, https://energy.gov/em/articles/moab-resumes-rail- shipments-after-rockslide (accessed November 2017).

181 Parker, J.M., 1981, Lisbon Field Area, San Juan County, Utah, in Wiegand, D.L., ed., Geology of the Paradox Basin, Denver, CO, Rocky Mountain Association of Geologists Field Conference, p. 89, http://archives.datapages.com/data/rmag/ParaBasin81/parker.htm (accessed November 2017).

Pevear, D.R., Vrolijk, P.J., and Longstaffe, F.J., 1997, Timing of Moab fault displacement and fluid movement integrated with burial history using radiogenic and stable isotopes: Geofluids II, v. 97, p. 42–45.

Richmond, G.M., 1962, Quaternary stratigraphy of the La Sal Mountains, Utah: U.S. Geological Survey Professional Paper 324, 135 p., http://pubs.er.usgs.gov/publication/pp324 (accessed October 2017).

Sharp, I.R., Gawthorpe, R.L., Underhill, J.R., and Gupta, S., 2000, Fault-propagation folding in extensional settings: Examples of structural style and synrift sedimentary response from the Suez rift, Sinai, Egypt: GSA Bulletin, v. 112, p. 1877–1899, doi: 10.1130/0016-7606(2000)112<1877:FPFIES>2.0.CO;2.

Solum, J.G., van der Pluijm, B.A., and Peacor, D.R., 2005, Neocrystallization, fabrics and age of clay minerals from an exposure of the Moab Fault, Utah: Journal of Structural Geology, v. 27, p. 1563–1576, doi: 10.1016/j.jsg.2005.05.002.

UDOGM, 1973, Well file of the Federal-Weaver No. 1 well, Grand County, Utah:, https://datamining.ogm.utah.gov/.

U.S. Geological Survey, and Utah Geological Survey, 2006, Quaternary fault and fold database for the United States: https://earthquake.usgs.gov/hazards/qfaults/ (accessed November 2017).

USGS Earthquake Hazards Program, https://earthquake.usgs.gov/ (accessed February 2018).

182 Weir, G.W., Puffett, W.P., and Dodson, C.L., 1994, Solution-collapse breccia pipes of Spanish Valley, southeastern Utah: U.S. Geological Survey Open-File Report 94– 417, 33 p., http://pubs.er.usgs.gov/publication/ofr94417 (accessed November 2017).

Williams, P.L., 1964, Geology, structure, and uranium deposits of the Moab quadrangle, Colorado and Utah: IMAP USGS Numbered Series 360, http://pubs.er.usgs.gov/publication/i360 (accessed February 2018).

Wong, I.G., and Humphrey, J.R., 1989, Contemporary seismicity, faulting, and the state of stress in the Colorado Plateau: GSA Bulletin, v. 101, p. 1127–1146, doi: 10.1130/0016-7606(1989)101<1127:CSFATS>2.3.CO;2.

Wong, I.G., Olig, S.S., and Bott, J.D.J., 1996, Earthquake Potential and Seismic Hazards in the Paradox Basin, Southeastern Utah, in Huffman, A.C.J., Lund, W.R., and Godwin, L.H., eds., Geology and Resources of the Paradox Basin, Salt Lake City, UT, Utah Geological Association Guidebook 25, p. 241–250, http://archives.datapages.com/data/uga/data/067/067001/241_ugs670241.htm (accessed November 2017).

Wong, I.G., and Simon, R.B., 1981, Low-Level Historical and Contemporary Seismicity in the Paradox Basin, Utah and its Tectonic Implications, in Wiegand, D.L., ed., Geology of the Paradox Basin, Denver, CO, Rocky Mountain Association of Geologists Field Conference, p. 169, http://archives.datapages.com/data/rmag/ParaBasin81/wong.htm (accessed November 2017).

Woodward-Clyde Consultants, 1986, Paradox Basin draft project summary report: Battelle Memorial Institute Chapter 1 Geology.

183 CHAPTER 5

SUMMARY

Chronostratigraphy of deformed alluvial deposits

Mapping and surveying along Mill and Pack creeks reveal five levels of terraces, two levels of upland gravels, and undifferentiated alluvial basin fill beneath the floor of

Moab and Spanish valleys. Luminescence and cosmogenic burial dating establish a chronostratigraphy for these alluvial deposits and provide insights into the timing, rates, and spatial patterns of fluvial incision and salt deformation. Terraces along Mill Creek are offset at the graben margin, and the T4 terrace correlates in age to exposed and down- faulted basin fill where the creek enters Moab Valley. The offset and age of these deposits constrains slip on the Kayenta Heights fault zone to 439 ± 49 m/my (0.439 ±

0.049 m/ky) in the past ~100 ky, while the maximum subsidence rate in the center of

Moab Valley may be as rapid as 1000 to 1350 m/my (1.00 to 1.35 m/ky). Terraces along

Pack Creek are down-warped into Spanish Valley as the drainage enters the graben axially. Projecting the tilted tread of the Pack Creek T3 terrace suggests that equivalent gravels are 45 m below grade in central Spanish Valley and implies a subsidence rate of

~450 m/my (0.45 m/ky) since ~100 ka. Given the landscape position and age of upland gravels that likely represent remnants of a relatively continuous early-Pleistocene piedmont apron recording baselevel stability, subsidence must have accelerated in the middle Pleistocene to the rapid rates recorded by terraces. A similar inflection is documented in incision rates in South Mill Creek, which accelerate from <50 m/my (0.05 m/ky) to >620 m/my (0.62 m/ky) prior to 200 ka. Post-200 ka incision rates near the

184 graben margin along Mill and Pack creeks are comparable at 529 ± 44 m/my (0.529 ±

0.044 m/ky) and 501 ± 65 m/my (0.501 ± 0.065 m/ky), respectively. The pattern of

terraces T3 through T1 in Mill Creek, which have straths that originate above-grade at upstream bedrock knickpoints and grow progressively higher downstream, suggests headward-migrating incision in this system. Together, these calculated rates and field observations provide evidence that late Pleistocene and Holocene transient incision is responsible for much of the canyon-cutting upstream of the Moab-Spanish Valley graben.

Acceleration of graben subsidence seems to have occurred at the same time as the

>200 ka increase in incision captured in the South Mill Creek chronostratigraphy. This

implies that salt-driven subsidence is coupled to baselevel fall along the Colorado River

through groundwater dissolution and transport, as has been proposed by previous workers

(e.g. Cater, 1970; Huntoon, 1982; Doelling et al., 1988; Gutiérrez, 2004). The baselevel

fall responsible for the transient incision in this system is not understood. It may relate to

the ~6 Ma integration of the Colorado River system off the Colorado Plateau or a

younger and yet-unknown source (Cook et al., 2009; Darling et al., 2012; Jochems and

Pederson, 2015). The process link between Colorado River incision and salt dissolution,

and the unknown baselevel source of the late-Pleistocene wave of incision, are important

future research topics to expand understanding of the landscape evolution of the central

Colorado Plateau. Nonetheless, from the chronostratigraphy presented here it is apparent

that both the Moab-Spanish Valley salt graben and the surrounding canyons are primarily

Quaternary features, adding to the growing body of evidence that the Moab region

represents a youthful landscape.

185 Topographic insights into baselevel, substrate control, and active salt tectonism

The central Colorado Plateau is a bedrock-dominated landscape where substrate resistance exerts a primary control on the form of stream longitudinal profiles. Colorado

River tributaries around the Moab-Spanish Valley salt graben are heavily influenced by the horizontal succession of stratigraphic units in this famous landscape. The juxtaposition of relatively resistant and weak lithology produces longitudinal profiles that do not conform to the equilibrium concave shape, instead exhibiting knickzones across resistant units like the Moab Member of the Curtis Formation. Longitudinal-profile indices of concavity (θ) and normalized steepness (ksn) reflect this bedrock control, with

high-θ reaches produced by upstream-resistant and downstream-erodible lithology and

high ksn values occurring in bedrock knickzones and headwater reaches where channels

are choked by Quaternary periglacial deposits. The planform and elevation distribution of broad knickzones closely follows outcrops of resistant lithologic units on the flanks of the

La Sal Mountains, though knickpoints in downstream bedrock canyons in weak, homogenous bedrock are interpreted as transient responses to baselevel fall according to strath terraces that emerge above-grade at their lips. When analyzed in Chi space, the largest knickzones in the study tributaries collapse to similar positions according to contributing area, suggesting they are derived from the same baselevel fall and are a product of transient incision. The baselevel source of this transient incision is the incising

Colorado River rather than a local effect of the subsiding salt graben, given that both graben-crossing and graben-exterior tributaries collapse on the Chi axis. Thus, though bedrock strength exerts the primary control, subtle baselevel signals from the Colorado

186 River are embedded in the longitudinal profiles of tributaries near the Moab-Spanish

Valley salt graben.

These results illustrate that profile analyses beyond traditional slope-area methods can detect baselevel signals in bedrock landscapes, and that large knickzones in the Moab region are the product of transient incision from past baselevel fall that is now focused and superimposed on resistant lithology. That transient incision from the Colorado River has propagated across the subsiding Moab-Spanish Valley salt graben and into the upper bedrock reaches of the study tributaries inspires a conceptual model whereby tributaries remain graded to the Colorado River despite ongoing graben subsidence. This salt- dissolution subsidence, which has occurred more rapidly than incision along the Colorado

River, is compensated by depositional filling of the basin with alluvium. This enables tributaries to maintain a connection to the regional baselevel of the Colorado River, lessening the influence of the subsiding graben as a local baselevel to upstream reaches.

Future studies of transient incision and baselevel signals on the Colorado Plateau would do well to incorporate both a thorough understanding of the bedrock template as well as integrative profile analysis such as Chi to completely address the multiple drivers of landscape evolution in this region.

Mapping of Spanish Valley

Mapping of the southern half of the Rill Creek and northern half of the Kane

Springs 7.5’ quadrangles informs patterns of deformation in and around Spanish Valley.

Mapped patterns of terraces along Pack Creek in southeast Spanish Valley confirms observations from previous workers (Richmond, 1962; Harden et al., 1985) and

187 documents that the T2 (Qatg2) and T3 (Qatg3) terrace treads project into the valley floor.

Bedrock mapping in the canyons and uplands outside of Spanish Valley exhibits early

Jurassic to late Cretaceous strata with a 5-10° southwest regional dip. These rocks are minimally deformed in the northeast and southwest corners of the map area and become

more heavily jointed and back-tilted in proximity of the graben margin. Structural mapping along the margin of Spanish Valley reveals that the graben-bounding deformation zone does not contain typical normal faults and that the deformation style differs between the northeast and southwest margins. In general, the fault zones along the graben margins are wide (>0.5 km), individual strands are short (average 0.7 km in length) with small (tens of meters) displacements, and are steeply dipping (up to 80°).

The northeast margin, which is part of the Kayenta Heights fault zone, is characterized by a series of two to three parallel, linear, down-to-the-southwest normal faults. These faults are exposed in the Navajo Sandstone footwall block where they accommodate a cumulative ~200 m of displacement, though there are likely additional strands concealed in the hangingwall block beneath the alluvium of Spanish Valley.

There is a lower density of faults along the northeast margin relative to the southwest margin and no mapped antithetic structures. Faults are parallel with pervasive northwest- trending joints and are commonly marked by cataclastic deformation bands, brecciation, and bleaching. The northeast margin also exhibits notable conical hills and linear ridges of fissure-fill breccia displaced from higher stratigraphic units that likely represent deposition of overburden into extensional fractures during prior episodes of deformation.

The age of these enigmatic, out-of-place, and topographically-inverted deposits is unknown.

188 The southwest graben margin is co-located with the southern section of the Moab fault zone and is typified by a series of master and subsidiary faults that accommodate strata of the Glen Canyon Group rolling over into the floor of the graben. Total

displacement across this fault zone is estimated at ~475 m. The master faults are down- to-the-northeast and the subsidiary faults have both synthetic and antithetic senses of slip.

The master faults record major offset of 200-250 m, are farthest from the graben axis, and

are responsible for the prominent escarpment southwest of Spanish Valley. Fault planes

dip southwest at lower-angles graben-ward and accommodate the progressive tilting of

bedding into the graben. Alluvial deposits onlap northeast-dipping bedrock flatirons at

the northeast edge of the fault zone, and there are likely additional concealed faults

beneath these basin sediments. Unlike the northeast graben margin, the southwest margin

lacks fissure-fill breccia deposits and cataclastic deformation bands.

Practical implications for the city of Moab

The renowned scenery of the Canyonlands region is a product of geologic

processes that are active today. These processes that continue to shape the landscape also

bear on the community of Moab, which sits in the alluvial valley floor of the Moab-

Spanish Valley graben. Slip across the Kayenta Heights fault zone as calculated at Mill

Creek has occurred at an average rate of 0.44 ± 0.05 mm/yr (44 ± 5 cm/ky) over the last

100 ky. This deformation is spatially and temporally heterogenous, and from prior studies

it is clear that local slip over shorter time intervals can greatly exceed these average rates

at up to 3 mm/yr, leading to potential surface rupture (Guerrero et al., 2015). Thus, this

vertical deformation can pose a hazard to structures built directly over a fault trace, which

189 is a common occurrence in subdivisions along the Kayenta Heights fault zone. Despite

this active surface deformation, the salt-dissolution faults along the northeast and southwest graben margin are not considered true tectonic faults and are not believed to accumulate and release strain in large seismic events (Wong and Humphrey, 1989; Olig et al., 1996; Wong et al,. 1996; Hylland and Mulvey, 2003). Instead, these steeply- dipping faults are likely shallowly-rooted failure planes that extend only hundreds of meters into the Paradox Formation (Cater, 1970; Doelling et al., 2002; Gutiérrez, 2004).

Though the up-to 1.00 to 1.35 mm/yr subsidence in the middle of Moab Valley is likely

not itself a hazard, acute dissolution of Paradox Formation evaporites where these rocks

are close to the surface can produce sinkhole openings, as occurred in 2015. The only

surface exposures of the Paradox Formation are along the flanks of Moab Valley

(Doelling et al., 2002), and development in these areas should be avoided.

Another hazard for the community of Moab are the mass-wasting processes along

the margins of Moab and Spanish valleys. Holocene rock fall and debris flow deposits are

mapped along both graben margins and are especially prevalent along the southwest

escarpment. As the valley floor continues to subside and cliffs along the graben margins

continue to retreat, rock falls, rock avalanches, and debris flows will be persistent

hazards.

Future development and zoning decisions in Moab and Spanish valleys would do

well to consider the geologic hazards of this active landscape, which include surface

rupture along salt-dissolution faults, acute dissolution of Paradox Formation evaporites,

and mass movements at the base of escarpments. Each of these hazards can be avoided by

not placing homes or other important infrastructure along the graben margins.

190 REFERENCES

Cater, F.W., 1970, Geology of the salt anticline region in southwestern Colorado: U.S. Geological Survey Professional Paper 637, 80 p., https://pubs.usgs.gov/pp/0637/report.pdf.

Doelling, H.H., Oviatt, C.G., and Huntoon, P.W., 1988, Salt deformation in the Paradox region: Utah Geological Survey, v. Bulletin 122.

Doelling, H.H., Ross, M.L., and Mulvey, W.E., 2002, Geologic Map of the Moab 7.5’ Quadrangle, Grand County, Utah: Utah Geological Survey.

Harden, D.R., Biggar, N.E., and Gillam, M.L., 1985, Quaternary deposits and soils in and around Spanish Valley, Utah: GSA Special Papers, v. 203, p. 43–64, doi: https://dx.doi.org/10.1130/SPE203.

191 Jochems, A.P., and Pederson, J.L., 2015, Active salt deformation and rapid, transient incision along the Colorado River near Moab, Utah: Journal of Geophysical Research: Earth Surface, v. 120, p. 2014JF003169, doi: 10.1002/2014JF003169.

Richmond, G.M., 1962, Quaternary stratigraphy of the La Sal Mountains, Utah: U.S. Geological Survey Professional Paper 324, 135 p., http://pubs.er.usgs.gov/publication/pp324 (accessed October 2017).

192

APPENDICES

193

Appendix A: Geochronology data

Luminescence data

TABLE A.1. FINAL LUMINESCENCE GEOCHRONOLOGY FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS Sample Latitude Longitude Elevation Depth Dose rate Equivalent dose OD Age ± 1σ Deposit* number (DD)† (DD)† (m) (m) Method§ Aliquots# (Gy/ky)** (Gy )** (%)†† (ka) Mill Creek T1 USU-2505 38.56304 -109.51630 1293 1.9 OSL-b 16 (23) 2.89 ± 0.13 40.06 ± 5.12 22.9 13.84 ± 1.60 T2 USU-2498 38.56169 -109.51772 1304 3.0 OSL-c 20 (26) 2.21 ± 0.10 125.87 ± 17.48 27.6 57.06 ± 6.81 T3 USU-2504 38.56425 -109.51315 1328 2.8 OSL-c 21 (37) 2.69 ± 0.12 217.93 ± 21.82 17.2 81.01 ± 8.81 T3 USU-2578 38.55297 -109.49403 1374 2.4 OSL-c 21 (28) 2.27 ± 0.10 186.80 ± 28.20 30.4 82.43 ± 10.07 T4y USU-2579 38.50890 -109.42657 1611 2.0 IRSL-b 12 (14) 4.66 ± 0.52 304.67 ± 69.62 38.3 93.47 ± 16.77 T5 USU-2506 38.55874 -109.51043 1395 1.9 IRSL-b 12 (15) 4.38 ± 0.48 565.94 ± 46.50 10.1 194.4 ± 27.2 ba, upper USU-2155 38.55885 -109.52442 1297 4.0 OSL-c 18 (28) 2.70 ± 0.12 286.90 ± 25.86 14.8 106.3 ± 11.3 ba, middle USU-1137§§ 38.55862 -109.52481 1283 10.0 OSL-c 13 (26) 1.26 ± 0.07 172.75 ± 22.95 20.3 136.8 ± 16.7 ba, basal USU-2497 38.55646 -109.52388 1284 20.4 IRSL-c 13 (13) 3.63 ± 0.47 375.49 ± 19.16 1.8 132.6 ± 20.6

Pack Creek T1y USU-2255 38.47260 -109.44423 1521 2.5 OSL-c 14 (15) 1.98 ± 0.09 11.79 ± 2.11 28.1 5.96 ± 0.78 T1 USU-2154 38.45943 -109.42851 1595 1.1 OSL-c 14 (16) 2.28 ± 0.11 36.30 ± 2.04 0.0 15.94 ± 1.58 T2 USU-2500 38.44848 -109.37690 1798 2.8 OSL-c 17 (23) 2.46 ± 0.11 163.76 ± 18.98 22.8 66.50 ± 7.45 T2 USU-2501 38.45156 -109.41365 1655 12.6 OSL-c 17 (32) 3.30 ± 0.15 258.47 ± 46.36 33.4 78.22 ± 10.36 gp, upper USU-2254 38.49062 -109.45771 1462 7.8 OSL-c 16 (22) 1.72 ± 0.12 199.67 ± 22.53 16.5 115.9 ± 14.1 gp, lower USU-2251 38.48957 -109.45903 1444 26.4 OSL-c 19 (22) 2.09 ± 0.21 226.15 ± 32.94 28.9 108.2 ± 16.2 *Organized by stratigraphic position; "T" stands for terrace deposit, "ba" for bowling alley site, and "gp" for gravel pit; ba and gp deposits are undifferentiated alluvium. †Coordinates reported in NAD83. §Abbreviations: b = bulk sample, c = core. #Number accepted, total aliquots in parentheses. **Reported with 2σ error. †† Overdispersion (OD) represents variance in equivalent dose beyond measurement uncertainties; included in equivalent dose error. 194 §§ Sample collected by F. Geiger in 2012; final number of accepted aliquots and age updated with rejection criteria from this study.

TABLE A.2. SUMMARY OF LUMINESCENCE DOSE RATE DATA FOR ALLUVIAL DEPOSITS OF MILL AND PACK CREEKS

Depth In-situ H2O Grain size K Rb Th U Cosmic dose rate Dose rate * † § § § § Deposit Sample number (m) (%) (µm) (%) (ppm) (ppm) (ppm) (Gy/ky) (Gy/ky) Mill Creek T1 USU-2505# 1.9 0.7 125-212 1.17±0.03 34.8±1.4 2.5±0.2 1.0±0.1 0.21±0.02 2.89 ± 0.13 (5) 2.47±0.06 57.7±2.3 4.8±0.4 2.7±0.3

2.10±0.05 57.3±2.3 6.1±0.5 2.1±0.2 T2 USU-2498** 3.0 0.6 125-212 1.84±0.05 62.4±2.5 2.7±0.2 0.9±0.1 0.18±0.02 2.21 ± 0.10 (5) 1.30±0.03 31.3±1.3 3.8±0.3 1.8±0.1 T3 USU-2504# 2.8 0.8 125-212 1.23±0.03 36.4±1.5 2.6±0.2 1.5±0.1 0.19±0.02 2.69 ± 0.12 (5) 2.19±0.05 46.4±1.9 4.6±0.4 2.3±0.2

2.04±0.05 43.6±1.7 5.3±0.5 2.0±0.2 T3 USU-2578# 2.4 1.2 90-250 1.43±0.04 45.3±1.8 2.4±0.2 0.8±0.1 0.26±0.03 2.27 ± 0.10 (5) 3.65±0.09 79.6±3.2 8.4±0.8 2.9±0.4

1.29±0.03 30.4±1.2 2.7±0.2 1.2±0.2 T4y USU-2579# 2.0 3 125-212 1.57±0.04 50.7±2.0 4.2±0.4 1.6±0.1 0.21±0.02 4.66 ± 0.52 (5) 1.91±0.05 57.6±2.3 4.9±0.4 1.9±0.1

2.27±0.06 53.7±2.1 5.8±0.5 2.0±0.1 T5 USU-2506# 1.9 2.2 125-212 1.42±0.04 48.5±1.9 3.4±0.3 1.3±0.1 0.21±0.02 4.38 ± 0.48 (5) 2.26±0.06 65.2±2.6 6.0±0.5 2.3±0.2

1.97±0.05 40.1±1.6 4.3±0.4 2.0±0.2 ba, upper USU-2155 4.0 1.4 90-150 1.79±0.04 67.4±2.7 6.1±0.6 1.8±0.1 0.16±0.02 2.70 ± 0.12 (5) ba, middle USU-1137 10.0 0 125-212 1.03±0.03 35.0±1.4 2.1±0.2 0.6±0.1 0.09±0.01 1.26 ± 0.07 (10) ba, basal USU-2497 20.4 0.1 90-150 1.53±0.04 54.5±2.2 6.3±0.6 2.0±0.1 0.04±0.00 3.63 ± 0.47

(20) 195

Depth In-situ H2O Grain size K Rb Th U Cosmic dose rate Dose rate * † § § § § Deposit Sample number (m) (%) (µm) (%) (ppm) (ppm) (ppm) (Gy/ky) (Gy/ky) Pack Creek T1y USU-2255 2.5 0.4 125-212 1.18±0.03 43.7±1.7 4.2±0.4 1.7±0.1 0.20±0.02 1.98 ± 0.09 (5) T1 USU-2154 1.1 1.6 125-180 1.26±0.03 45.6±1.8 5.5±0.5 2.1±0.2 0.25±0.02 2.28 ± 0.11 (5) T2 USU-2500# 2.8 0.7 150-250 1.37±0.03 50.6±2.0 4.5±0.4 1.8±0.1 0.21±0.02 2.46 ± 0.11 (5) 1.42±0.04 37.8±1.5 5.8±0.5 2.8±0.2

1.68±0.04 36.3±1.5 5.5±0.5 1.7±0.1 T2 USU-2501** 12.6 0.8 125-212 1.18±0.03 44.2±1.8 5.3±0.5 2.2±0.2 0.07±0.01 3.30 ± 0.15 (5) 2.76±0.07 66.2±2.6 6.4±0.6 3.1±0.2 gp, upper USU-2254 7.8 3.1 125-212 1.09±0.03 40.9±1.6 4.8±0.4 2.0±0.1 0.11±0.01 1.72 ± 0.12 (15) gp, lower USU-2251 26.4 10.1 90-150 1.70±0.04 69.7±2.8 6.1±0.6 2.0±0.1 0.06±0.01 2.09 ± 0.21 (25)

*Organized by stratigraphic position; "T" stands for terrace deposit, "ba" for bowling alley site, and "gp" for gravel pit; ba and gp deposits are undifferentiated alluvium. †Value in parentheses used for moisture content over burial history in dose rate calculation. §Radioelemental concentrations determined by ALS Chemex using ICP-MS and ICP-AES techniques. #Beta and gamma DR uses weighted average of 3 g.s. fractions. Top value: sand including silt/clay, middle value: pebbles, bottom value: cobble. **Beta and gamma DR uses weighted average of 2 g.s. fractions. Top value: sand including silt/clay, bottom value: pebbles. 196

TABLE A.3. GRAIN SIZE PROPORTIONS AND α, β AND γ DOSE RATE INFORMATION

Sand Pebbles Cobbles 0.001-1.7mm 1.7-16mm >16mm α β γ Dose rate Deposit* Sample number (%)† (%)† (%)† (Gy/ky) (Gy/ky) (Gy/ky)§ (Gy/ky) Mill Creek T1 USU-2505 20 30 50 N.A.# 1.77 1.12 2.89 ± 0.13 T2 USU-2498 60 40 0 N.A.# 1.37 0.84 2.21 ± 0.10 T3 USU-2504 30 55 15 N.A.# 1.66 1.03 2.69 ± 0.12 T3 USU-2578 10 10 80 N.A.# 1.35 0.92 2.27 ± 0.10 T4y USU-2579 30 15 55 1.26 2.28 1.12 4.66 ± 0.52 USU-2506 35 25 40 1.18 2.15 1.05 4.38 ± 0.48 T5

Pack Creek T2 USU-2500 15 5 80 N.A.# 1.44 1.02 2.46 ± 0.11 T2 USU-2501 30 70 0 N.A.# 2.1 1.2 3.30 ± 0.15 *Organized by stratigraphic position. †Measurements rounded to nearest 5%. §Includes cosmic dose rate. #N.A. = not applicable.

197

198 OSL probability density functions and radial plots for equivalent dose distribution

USU-1137 D E

(Gy)

USU-2154 D E

(Gy)

USU-2155 D E

(Gy)

Figure A.1. continues

199 USU-2251 D E

(Gy)

USU-2254 D E

(Gy)

USU-2255 D E

(Gy)

Figure A.1. continues

200 USU-2498 D E

(Gy)

USU-2500 D E

(Gy)

USU-2501 D E

(Gy)

Figure A.1. continues

201 USU-2504 D E

(Gy)

USU-2505 D E

(Gy)

USU-2578 D E

(Gy)

Figure A.1. OSL probability density functions and radial plots for equivalent dose distribution.

202 IRSL probability density functions and radial plots for equivalent dose distribution

USU-2497 IRSL Age (ka) Age IRSL

USU-2506 IRSL Age (ka) Age IRSL

USU-2579 IRSL Age (ka) Age IRSL

Figure A.2. IRSL probabililty density functions and radial plots for equivalent dose distribution.

203 Cosmogenic radionuclide (CRN) data

TABLE A.4. PRELIMINARY COSMOGENIC RADIONUCLIDE GEOCHRONOLOGY FOR T7 DEPOSITS Source Latitude Longitude Elevation Depth elevation Age ± 1σ Sample site (DD) (DD) (m) (m) (m) (Ma)* Johnson's Ridge (JWUOT) 38.48878 -109.41846 1770 17 2300 ± 500 1.58 ± 0.50 South Mountain (HPM) 38.43456 -109.39624 1857 26 2700 ± 300 1.73 ± 0.26 *Calculated from isochron using Granger (2013) linearization method.

TABLE A.5. CRN SAMPLE SITE INFORMATION Sampling horizon Sample Latitude Longitude Elevation Depth thickness number (DD) (DD) (m) (m)* (m) Lithology JWUOT-01 38.48878 -109.41846 1770 17 0.7 Sand JWUOT-02 38.48878 -109.41846 1770 17 0.7 Salt and pepper sandstone JWUOT-03 38.48878 -109.41846 1770 17 0.7 Entrada Sandstone JWUOT-05 38.48878 -109.41846 1770 17 0.7 Gray fluvial sandstone HPM-03 38.43456 -109.39624 1857 26 3.1 Navajo Sandstone HPM-04 38.43456 -109.39624 1857 26 3.1 Quartz-rich sandstone HPM-08 38.43456 -109.39624 1857 26 3.1 Pink sandstone *Measured at top of samping horizon.

TABLE A.6. CRN AMS DATA 10Be 26Al Qtz count 10Be conc. 10Be total count 26Al conc. 26Al total Sample mass (106 (106 1σ error (106 (106 1σ error number (g) atoms) atoms/g) (%) atoms) atoms/g) (%) JWUOT-01 34.70 2.987 0.0861 1.92 14.427 0.4158 8.56 JWUOT-02 15.27 1.331 0.0872 1.80 7.301 0.4781 6.95 JWUOT-03 27.84 3.173 0.1140 1.70 12.148 0.4364 8.64 JWUOT-05 38.82 1.382 0.0356 1.80 9.285 0.2392 7.18 HPM-03 35.45 3.908 0.1102 1.70 16.971 0.4787 4.99 HPM-04 35.32 7.577 0.2145 2.75 26.403 0.7475 6.86 HPM-08 36.97 3.707 0.1003 1.70 14.835 0.4013 7.36

204

Figure A.3. Johnson’s Mesa CRN isochron. Output from Granger (2013) linearization code.

Figure A.4. South Mountain CRN isochron. Output from Granger (2013) linearization code.

205

Appendix B. GPS survey data

206 GPS survey details

Topographic survey transects were performed with the combination of a handheld, mapping-grade Trimble GeoXT GPS and a real-time kinematic (RTK), survey- grade TopCon HiPer GA GPS. The handheld unit was used for transects without reasonable road access to establish a RTK base station, and data were post-process differentially-corrected using the MONTICELLOUT2006 CORS base station. RTK surveys were performed by placing a stationary base receiver at a centrally-located high point within view of the survey. The base station auto-position was determined after ~30 minutes of signal reception, as there were no ground-control benchmarks available in the transect lines. After this period, survey points were collected with a mobile rover receiver, which receives real-time drift corrections from the base station. RTK elevations are consistently 20-26 m lower than elevations from the handheld GPS and the 10 m digital elevation model from the USGS National Elevation Dataset. This is likely due to the absence of ground-control points and because base stations were not allowed to auto- position for the recommended >4 hours, which is longer than the battery life of the RTK. For this reason, a single adjustment was applied to the raw elevation data from the RTK (“adjusted elevation” in tables) for each transect based off of manual comparison in GIS after the transects were completed. Handheld GPS data did not require an elevation adjustment. Transects were performed by collecting GPS points along a line perpendicular (or in the sole case of Mill2, parallel) to the valley. Points were collected at pertinent geomorphic positions (i.e. terrace straths and treads) and at inflections in topography. For both the RTK and handheld, the GPS was allowed to acquire 5-10 positions, which were aggregated into a single point coordinate. All coordinate and elevation data is in reference to the UTM Zone 12N projection, NAD83 datum, and g2012au2 geoid. In all transect graphs below, red dots show points outside of the main transect plane.

Figure B.1. GPS transect locations. Boxes and letters at left are keyed to large-scale views at right. Circles show transect points. 207

208

Figure B.2. Mill1 GPS transect (RK 5.0). Cross section, view downstream.

TABLE B.1. SURVEY DATA FOR MILL1 GPS TRANSECT (RK 5.0) Survey type: RTK Raw Adjusted Height Distance UTM 12N UTM 12N elevation elevation above along northing easting ASL ASL grade transect Point (m) (m) (m) (m)* (m) (m) Notes base 4268852.00 628586.69 1279.72 1299.72 40.24 N.A. base 1 4268752.08 628462.48 1274.20 1294.20 34.73 0.00 basin-fill tread 2 4268766.16 628456.64 1274.12 1294.12 34.64 15.24 basin-fill tread 3 4268780.35 628449.71 1273.88 1293.88 34.40 31.02 basin-fill tread 4 4268787.65 628445.86 1273.22 1293.22 33.75 39.25 basin-fill tread 5 4268794.81 628441.55 1270.53 1290.53 31.05 47.56 basin-fill riser 6 4268804.99 628434.19 1264.64 1284.64 25.17 59.95 basin-fill riser 7 4268817.18 628426.47 1257.32 1277.32 17.84 74.33 basin-fill riser 8 4268825.90 628420.28 1252.32 1272.32 12.84 84.91 basin-fill riser 9 4268830.39 628414.78 1250.65 1270.65 11.17 91.50 basin-fill riser 10 4268843.15 628409.07 1249.78 1269.78 10.30 105.37 T1 tread 11 4268882.40 628389.97 1248.53 1268.53 9.06 148.81 T1 tread 12 4268935.91 628363.72 1247.05 1267.05 7.57 208.12 T1 tread 13 4269030.65 628318.39 1246.01 1266.01 6.53 312.47 T1 tread 14 4269083.67 628295.82 1245.05 1265.05 5.57 369.66 T1 tread 15 4269137.48 628271.25 1244.12 1264.12 4.64 428.22 T1 tread 16 4269150.84 628265.51 1243.49 1263.49 4.01 442.61 T1 tread 17 4269168.18 628258.03 1243.29 1263.29 3.81 461.28 T1 tread 18 4269175.93 628254.07 1242.77 1262.77 3.30 469.82 T1 tread 19 4269176.99 628253.62 1242.36 1262.36 2.88 470.95 T1 riser 20 4269178.62 628253.15 1241.96 1261.96 2.49 472.63 T1y tread 21 4269204.91 628248.28 1242.19 1262.19 2.71 498.81 T1y tread 22 4269208.03 628247.56 1241.70 1261.70 2.22 501.97 T1y riser 23 4269212.03 628247.77 1241.34 1261.34 1.86 505.66 T1yy tread 24 4269214.92 628247.54 1241.57 1261.57 2.10 508.45 T1yy tread 25 4269216.54 628247.07 1241.38 1261.38 1.91 510.12 T1yy tread 26 4269218.09 628247.28 1240.86 1260.86 1.38 511.51 T1yy riser 27 4269219.15 628245.94 1239.93 1259.93 0.45 512.94 water surface 28 4269221.75 628245.84 1239.48 1259.48 0.00 515.42 channel bottom

209 29 4269224.39 628243.94 1239.87 1259.87 0.39 518.50 water surface 30 4269225.06 628245.20 1240.66 1260.66 1.19 518.73 T1yy riser 31 4269227.60 628245.64 1241.52 1261.52 2.04 520.98 T1yy tread 32 4269228.89 628245.71 1241.72 1261.72 2.24 522.18 T1yy tread 33 4269230.79 628245.99 1242.44 1262.44 2.96 523.87 T1 riser 34 4269236.12 628245.22 1243.50 1263.50 4.02 529.14 T1 tread 35 4269260.70 628195.26 1242.83 1262.83 3.35 566.13 T1 tread 36 4269261.37 628191.96 1242.00 1262.00 2.52 567.50 T1Y riser 37 4269262.25 628186.34 1241.26 1261.26 1.78 569.52 T1Y tread 38 4269308.04 628098.40 1239.09 1259.09 -0.38 618.25 T1Y tread 39 4269315.08 628097.60 1240.59 1260.59 1.12 624.27 basin-fill riser 40 4269319.38 628099.04 1242.80 1262.80 3.32 628.10 basin-fill riser 41 4269324.85 628099.20 1243.53 1263.53 4.05 632.85 basin-fill riser 42 4269330.54 628099.24 1245.59 1265.59 6.11 637.78 basin-fill riser 43 4269337.09 628098.38 1247.88 1267.88 8.40 643.35 basin-fill riser 44 4269346.35 628096.89 1251.98 1271.98 12.50 651.15 basin-fill riser 45 4269354.07 628097.83 1255.76 1275.76 16.28 657.92 basin-fill riser 46 4269364.99 628107.10 1260.04 1280.04 20.57 668.83 basin-fill riser 47 4269370.50 628107.39 1261.51 1281.51 22.03 673.71 basin-fill riser 48 4269380.50 628108.48 1262.97 1282.97 23.50 682.67 basin-fill riser 49 4269408.98 628118.24 1265.90 1285.90 26.43 709.78 basin-fill tread 50 4269443.36 628113.76 1266.10 1286.10 26.63 738.26 basin-fill tread 51 4269478.32 628125.06 1266.38 1286.38 26.90 771.53 basin-fill tread 52 4269491.47 628127.33 1266.99 1286.99 27.51 782.71 bedrock 53 4269531.21 628163.50 1271.86 1291.86 32.38 834.52 bedrock *20m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see Appendix B introduction).

210

Figure B.3. Mill2 GPS transect (RK 5.7-5.0). Survey is south of and parallel to Mill Creek, flowing left to right. KHFZ = Kayenta Heights fault zone.

TABLE B.2. SURVEY DATA FOR MILL2 GPS TRANSECT (RK 5.7-5.0) Survey type: RTK Raw Adjusted Height Distance elevation elevation above along Northing Easting ASL ASL grade transect Point (m) (m) (m) (m)* (m) (m) Notes base 4268852.00 628586.69 1279.72 1299.72 40.24 N.A. base 1 4268778.04 628301.92 1249.31 1269.31 9.83 494.12 T1 tread 2 4268786.23 628324.88 1250.49 1270.49 11.02 471.33 T1 tread 4 4268766.65 628342.32 1252.73 1272.73 13.25 460.14 basin-fill riser 5 4268766.49 628351.75 1258.99 1278.99 19.51 450.76 basin-fill riser 6 4268767.88 628354.41 1261.73 1281.73 22.25 448.00 basin-fill riser 7 4268765.92 628365.55 1265.81 1285.81 26.33 437.00 basin-fill riser 8 4268765.35 628375.22 1268.89 1288.89 29.41 427.37 basin-fill riser 9 4268765.62 628383.70 1271.33 1291.33 31.86 418.88 basin-fill riser 10 4268765.66 628389.52 1272.84 1292.84 33.36 413.06 basin-fill riser 11 4268765.74 628400.95 1274.60 1294.60 35.12 401.63 basin-fill tread 12 4268760.74 628413.45 1274.96 1294.96 35.49 389.18 basin-fill tread 13 4268760.51 628429.86 1274.73 1294.73 35.25 372.78 basin-fill tread 14 4268762.32 628457.16 1274.12 1294.12 34.65 345.48 basin-fill tread 15 4268762.85 628474.29 1274.02 1294.02 34.54 328.35 basin-fill tread 16 4268760.04 628493.29 1274.02 1294.02 34.54 309.36 basin-fill tread 17 4268756.47 628516.68 1274.20 1294.20 34.72 285.99 basin-fill tread 18 4268756.11 628548.14 1275.07 1295.07 35.59 254.56 basin-fill tread 19 4268756.22 628565.30 1275.92 1295.92 36.44 237.41 basin-fill tread 20 4268759.33 628644.64 1280.22 1300.22 40.74 158.10 basin-fill tread 21 4268760.83 628668.02 1281.51 1301.51 42.04 134.70 basin-fill tread 22 4268763.47 628699.75 1284.02 1304.02 44.54 102.92 basin-fill tread 23 4268766.78 628712.31 1285.49 1305.49 46.02 90.30 basin-fill tread 24 4268767.87 628724.35 1286.58 1306.58 47.11 78.25 basin-fill tread 25 4268769.89 628736.84 1287.36 1307.36 47.88 65.75 basin-fill tread

211 26 4268771.31 628742.85 1287.25 1307.25 47.77 59.76 basin-fill tread 27 4268770.87 628749.66 1286.59 1306.59 47.11 52.94 basin-fill tread 28 4268773.13 628753.64 1286.08 1306.08 46.60 49.01 bedrock 29 4268774.02 628756.58 1284.93 1304.93 45.45 46.11 bedrock 30 4268775.58 628766.16 1285.75 1305.75 46.27 36.62 bedrock 31 4268785.49 628781.18 1288.17 1308.17 48.69 22.91 bedrock 32 4268794.45 628798.44 1293.63 1313.63 54.15 8.80 bedrock 33 4268794.98 628804.93 1298.77 1318.77 59.30 3.13 bedrock 34 4268806.14 628815.28 1299.96 1319.96 60.48 0.00 bedrock 35 4269047.63 628736.70 1264.93 1284.93 25.45 152.49 T2 strath 36 4269052.17 628742.31 1264.91 1284.91 25.43 151.60 T2 strath 37 4269050.76 628750.42 1265.20 1285.20 25.72 148.28 T2 strath 38 4269042.25 628744.76 1267.56 1287.56 28.08 147.97 T2 tread 39 4269043.93 628737.40 1265.30 1285.30 25.82 151.23 T2 strath 40 4269035.22 628713.24 1266.29 1286.29 26.82 158.85 T2 tread *20m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see Appendix B introduction).

212

Figure B.4. Mill3 GPS transect (RK 6.3). Cross section, view downstream.

TABLE B.3. SURVEY DATA FOR MILL3 GPS TRANSECT (RK 6.3) Survey type: RTK Raw Adjusted Height Distance elevation elevation above along Northing Easting ASL ASL grade transect Point (m) (m) (m) (m)* (m) (m) Notes base 4269385.75 629185.26 1306.55 1326.55 41.88 N.A. base 1 4269116.53 629206.85 1293.25 1313.25 28.58 0.00 bedrock 2 4269119.94 629209.32 1291.50 1311.50 26.83 4.22 colluvium 3 4269124.04 629212.00 1290.07 1310.07 25.40 9.12 colluvium 4 4269132.43 629217.96 1288.20 1308.20 23.53 19.40 colluvium 5 4269135.57 629220.56 1287.61 1307.61 22.94 23.46 T2 tread 6 4269144.20 629227.22 1287.56 1307.56 22.88 34.36 T2 tread 7 4269146.68 629229.09 1287.25 1307.25 22.58 37.47 T2 tread 8 4269166.26 629188.95 1285.41 1305.41 20.73 50.71 T2 tread 9 4269167.34 629190.01 1285.15 1305.15 20.47 51.43 T2 riser 10 4269165.51 629193.46 1284.64 1304.64 19.97 49.08 T2 strath 11 4269168.62 629194.71 1284.20 1304.20 19.53 51.74 bedrock 12 4269171.07 629197.81 1281.95 1301.95 17.28 53.66 bedrock 13 4269175.71 629200.44 1277.00 1297.00 12.33 57.89 bedrock 14 4269172.82 629243.72 1277.10 1297.10 12.43 67.29 colluvium 15 4269179.86 629248.38 1276.32 1296.32 11.65 75.73 T1 tread 16 4269187.51 629252.97 1276.05 1296.05 11.38 84.64 T1 tread 17 4269193.56 629256.46 1276.97 1296.97 12.30 91.61 T1 tread 18 4269211.47 629265.18 1276.33 1296.33 11.66 111.38 T1 tread 19 4269223.90 629275.28 1275.93 1295.93 11.26 127.29 T1 tread 20 4269226.34 629279.27 1275.43 1295.43 10.76 131.53 T1 riser 21 4269228.81 629280.65 1274.95 1294.95 10.27 134.35 T1y tread 22 4269238.05 629286.46 1274.96 1294.96 10.29 145.25 T1y tread 23 4269242.39 629290.43 1274.06 1294.06 9.39 151.07 T1y tread 24 4269244.77 629292.00 1274.41 1294.41 9.74 153.92 T1y tread 25 4269254.04 629298.78 1273.82 1293.82 9.15 165.40 T1y tread 26 4269259.47 629302.53 1273.32 1293.32 8.65 172.00 T1y tread 27 4269254.13 629337.84 1272.85 1292.85 8.18 188.96 T1y strath

213 28 4269243.11 629368.78 1273.84 1293.84 9.17 200.78 T1 strath 29 4269258.27 629338.53 1272.37 1292.37 7.69 192.57 bedrock 30 4269262.20 629340.29 1271.18 1291.18 6.51 196.70 bedrock 31 4269264.72 629342.07 1270.16 1290.16 5.48 199.76 bedrock 32 4269267.27 629342.84 1266.41 1286.41 1.74 202.21 colluvium 33 4269267.04 629350.32 1265.88 1285.88 1.21 206.65 bedrock 34 4269269.07 629351.59 1265.19 1285.19 0.52 209.00 water surf 35 4269269.52 629354.08 1264.67 1284.67 0.00 210.88 channel bottom 36 4269270.91 629354.28 1264.93 1284.93 0.26 212.08 water surf 37 4269272.79 629354.47 1266.66 1286.66 1.99 213.64 colluvium 38 4269279.79 629345.41 1267.20 1287.20 2.53 213.62 bedrock 39 4269284.08 629346.93 1268.75 1288.75 4.08 217.93 bedrock 40 4269285.13 629347.00 1270.42 1290.42 5.74 218.81 bedrock 41 4269287.86 629348.27 1271.52 1291.52 6.85 221.74 bedrock 42 4269288.17 629349.54 1273.28 1293.28 8.61 222.75 bedrock 43 4269291.64 629351.61 1274.00 1294.00 9.33 226.75 T1 strath 44 4269295.72 629325.90 1274.24 1294.24 9.57 215.09 T1 strath 45 4269296.98 629326.74 1274.69 1294.69 10.02 216.61 T1 riser 46 4269299.26 629328.14 1275.18 1295.18 10.50 219.28 T1 tread 47 4269304.33 629329.98 1275.36 1295.36 10.69 224.49 T1 tread 48 4269313.44 629336.15 1275.93 1295.93 11.25 235.49 T1 tread 49 4269325.58 629346.60 1276.27 1296.27 11.60 251.41 T1 tread 50 4269336.01 629354.63 1276.90 1296.90 12.22 264.55 T1 tread 51 4269341.48 629354.49 1276.97 1296.97 12.30 268.95 bedrock 52 4269349.86 629354.65 1277.87 1297.87 13.20 275.90 bedrock 53 4269351.26 629376.41 1279.03 1299.03 14.36 289.54 bedrock 54 4269356.45 629379.42 1283.29 1303.29 18.62 295.53 bedrock 55 4269356.85 629384.04 1286.56 1306.56 21.89 298.50 bedrock 56 4269359.01 629388.67 1288.60 1308.60 23.93 302.92 T2 strath 57 4269360.45 629390.77 1289.37 1309.37 24.70 305.31 T2 riser 58 4269362.49 629394.68 1290.39 1310.39 25.72 309.21 T2 tread 59 4269369.52 629398.72 1290.67 1310.67 25.99 317.27 T2 tread 60 4269393.95 629392.26 1290.52 1310.52 25.84 333.54 colluvium 61 4269392.51 629395.01 1291.48 1311.48 26.81 333.97 T2o strath 62 4269392.41 629398.47 1292.40 1312.40 27.73 335.89 T2o riser 63 4269391.63 629402.78 1293.07 1313.07 28.40 337.74 T2o tread 64 4269397.68 629417.05 1293.53 1313.53 28.86 350.81 T2o tread 65 4269403.61 629425.03 1294.93 1314.93 30.26 360.17 slopewash 66 4269411.97 629430.42 1296.14 1316.14 31.47 370.09 slopewash 67 4269420.84 629437.07 1298.34 1318.34 33.67 381.13 colluvium 68 4269429.06 629441.50 1300.49 1320.49 35.82 390.41 colluvium 69 4269435.06 629445.99 1302.65 1322.65 37.98 397.88 colluvium 70 4269440.23 629449.17 1304.27 1324.27 39.60 403.93 colluvium 71 4269445.17 629452.40 1305.73 1325.73 41.06 409.82 T3 riser 72 4269450.51 629455.99 1306.97 1326.97 42.30 416.25 T3 riser 73 4269457.03 629461.96 1307.90 1327.90 43.23 424.92 T3 tread 74 4269464.73 629467.62 1308.07 1328.07 43.40 434.42 T3 tread 75 4269472.82 629470.57 1307.08 1327.08 42.41 442.83 T3 riser 76 4269481.62 629476.58 1305.19 1325.19 40.51 453.46 T3 riser

214 77 4269496.84 629486.06 1303.34 1323.34 38.67 471.34 slopewash 78 4269549.20 629515.94 1305.73 1325.73 41.06 531.55 bedrock 79 4269584.01 629535.74 1307.61 1327.61 42.94 571.59 bedrock 80 4269392.19 629577.82 1292.08 1312.08 27.41 401.14 bedrock 81 4269395.77 629581.73 1292.95 1312.95 28.27 404.08 T2o strath 82 4269398.01 629581.32 1293.69 1313.69 29.02 405.73 T2o riser 83 4269400.22 629581.99 1294.55 1314.55 29.88 407.42 T2o tread 84 4269403.23 629583.28 1295.12 1315.12 30.45 409.73 T2o tread 85 4269407.94 629584.89 1295.68 1315.68 31.01 413.30 T2o tread 86 4269411.59 629585.74 1296.56 1316.56 31.89 416.05 colluvium 87 4269416.94 629586.36 1298.41 1318.41 33.74 420.08 colluvium 88 4269424.25 629586.79 1301.28 1321.28 36.61 425.59 colluvium 89 4269431.52 629590.03 1304.07 1324.07 39.39 430.96 colluvium 90 4269442.72 629595.29 1308.12 1328.12 43.45 438.92 T3 strath 91 4269447.26 629598.13 1310.18 1330.18 45.51 441.93 T3 riser 92 4269451.51 629599.52 1311.75 1331.75 47.08 444.85 T3 riser 93 4269456.09 629600.62 1312.74 1332.74 48.07 448.01 T3 tread 94 4269460.68 629600.43 1312.53 1332.53 47.86 451.35 bedrock 95 4269469.24 629604.12 1311.21 1331.21 46.54 456.73 bedrock 96 4269485.19 629616.88 1309.23 1329.23 44.56 463.99 bedrock 97 4269490.04 629621.44 1310.83 1330.83 46.16 465.50 bedrock 98 4269466.18 629541.04 1306.97 1326.97 42.30 459.52 T3 strath *20m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see Appendix B introduction)

215

Figure B.5. Mill4 GPS transect (RK 9.3). Cross section, view downstream

TABLE B.4. SURVEY DATA FOR MILL4 GPS TRANSECT (RK 9.3) Survey type: handheld Height Distance Elevation above along Northing Easting ASL grade transect Point (m) (m) (m) (m) (m) Notes 1 4268206.91 631084.93 1413.66 70.23 0.00 eolian sand 2 4268210.31 631092.51 1412.85 69.43 8.31 bedrock 3 4268224.50 631107.42 1409.16 65.73 28.53 bedrock 4 4268231.07 631118.21 1402.70 59.28 41.11 bedrock 5 4268242.05 631125.34 1394.47 51.05 53.41 eolian sand 6 4268252.44 631148.36 1386.35 42.93 78.02 eolian sand 7 4268265.33 631169.94 1379.83 36.41 103.09 eolian sand 8 4268277.41 631193.07 1375.23 31.81 129.06 eolian sand 9 4268278.96 631202.11 1374.54 31.12 137.56 T3 tread 10 4268280.76 631209.64 1374.63 31.21 144.91 T3 tread 11 4268282.48 631216.37 1373.14 29.72 151.55 T3 riser 12 4268283.17 631219.30 1372.44 29.02 154.41 T3 riser 13 4268284.37 631220.06 1371.92 28.50 155.68 T3 strath 14 4268286.24 631224.19 1369.49 26.07 160.18 bedrock 15 4268286.84 631226.74 1368.98 25.56 162.66 bedrock 16 4268292.87 631231.70 1350.98 7.56 170.05 sandy alluvium 17 4268299.35 631238.96 1352.89 9.47 179.62 sandy alluvium 18 4268303.42 631244.31 1353.39 9.97 186.32 sandy alluvium 19 4268310.05 631245.38 1352.37 8.95 190.70 T2-T1 tread 20 4268313.15 631246.40 1349.36 5.94 193.17 T2-T1 riser 21 4268314.83 631248.09 1350.21 6.79 195.48 T2-T1 strath 22 4268317.39 631249.49 1349.60 6.17 197.98 water surface 23 4268322.02 631253.87 1350.23 6.81 204.08 T2-T1 strath

216 24 4268310.57 631277.93 1351.74 8.32 218.28 T2-T1 tread 25 4268328.96 631258.28 1351.13 7.71 211.30 sandy alluvium 26 4268331.00 631259.19 1352.31 8.89 213.06 sandy alluvium 27 4268340.26 631268.92 1352.06 8.64 225.83 sandy alluvium 28 4268345.81 631285.80 1355.92 12.50 243.25 sandy alluvium 29 4268347.46 631287.93 1357.51 14.09 245.86 sandy alluvium 30 4268347.76 631289.38 1362.37 18.95 247.29 bedrock 31 4268352.87 631315.51 1367.29 23.87 272.81 bedrock 32 4268353.61 631334.57 1369.33 25.91 289.04 bedrock 33 4268363.98 631355.86 1369.57 26.15 310.77 bedrock 34 4268378.76 631363.41 1370.07 26.65 327.23 bedrock 35 4268510.69 631103.87 1366.48 23.06 241.36 bedrock 36 4268516.65 631110.12 1369.85 26.42 249.28 bedrock 37 4268523.14 631113.60 1374.14 30.71 256.04 bedrock 38 4268522.77 631117.71 1375.31 31.89 257.91 T3 strath 39 4268528.21 631114.03 1375.61 32.19 260.12 T3 tread 40 4268535.36 631109.29 1374.97 31.55 262.36 T3 strath 41 4268540.90 631099.39 1372.92 29.50 258.26 T3 strath 42 4268542.27 631093.10 1370.49 27.07 253.84 T3 strath 43 4268534.48 631114.85 1372.63 29.21 265.40 bedrock 44 4268541.75 631120.35 1369.08 25.66 274.33 colluvium 46 4268561.54 631133.31 1365.40 21.98 297.79 bedrock 47 4268520.99 631089.64 1356.09 12.66 238.25 colluvium 48 4268505.22 631077.60 1345.66 2.24 219.00 colluvium 49 4268503.40 631075.57 1344.70 1.28 216.31 bedrock 50 4268502.32 631073.58 1344.49 1.06 214.09 bedrock 51 4268494.81 631063.88 1344.63 1.21 201.95 bedrock 52 4268494.18 631062.58 1345.32 1.90 200.53 water surface 53 4268493.52 631059.88 1343.42 0.00 197.98 water surface 54 4268490.42 631053.33 1345.88 2.45 190.89 sandy alluvium 55 4268487.45 631050.67 1347.22 3.79 187.01 sandy alluvium 56 4268481.88 631035.31 1349.24 5.82 171.65 eolian sand 57 4268470.95 631025.56 1354.73 11.31 157.01 colluvium

217

Figure B.6. Mill5 GPS transect (RK 19.0). Cross section, view downstream.

TABLE B.5. SURVEY DATA FOR MILL5 GPS TRANSECT (RK 19.0) Survey type: handheld Height Distance Elevation above along Northing Easting ASL grade transect Point (m) (m) (m) (m) (m) Notes 1 4262970.26 636235.87 1656.83 100.58 0.00 bedrock 2 4262981.40 636243.28 1658.92 102.67 10.80 colluvium 3 4262985.99 636256.11 1661.70 105.45 17.04 colluvium 4 4262987.11 636270.80 1663.33 107.08 22.07 colluvium 5 4262986.07 636279.50 1665.47 109.22 24.36 colluvium 6 4262985.61 636293.13 1667.79 111.54 30.16 colluvium 7 4262988.10 636311.49 1672.07 115.82 42.32 colluvium 8 4262993.86 636329.55 1676.76 120.51 58.11 colluvium 9 4262997.94 636341.00 1679.74 123.49 68.80 colluvium 10 4263001.16 636353.54 1682.98 126.73 79.90 colluvium 11 4263007.35 636366.19 1687.16 130.91 93.32 colluvium 12 4263012.32 636377.14 1691.33 135.08 104.76 colluvium 13 4263021.18 636388.30 1695.13 138.88 118.93 colluvium

218 14 4263029.06 636402.71 1699.94 143.68 134.95 colluvium 15 4263032.05 636410.21 1702.23 145.98 142.63 bedrock 16 4263036.64 636417.60 1704.97 148.71 151.24 bedrock 17 4263047.46 636425.69 1710.47 154.21 164.27 bedrock 18 4263051.33 636435.84 1716.89 160.64 174.56 T7 strath 19 4263016.13 636456.80 1715.31 159.06 171.67 T7 strath 20 4262970.81 636508.56 1717.62 161.37 197.80 T7 strath 21 4263061.17 636432.31 1715.65 159.40 178.01 T7 strath 22 4263080.97 636414.29 1713.62 157.36 177.92 T7 strath 23 4263053.82 636438.78 1718.65 162.40 178.40 T7 riser 24 4263054.50 636447.23 1722.55 166.29 185.47 T7 riser 25 4263056.90 636456.97 1725.94 169.68 194.67 T7 riser 26 4263060.44 636466.95 1729.47 173.21 204.79 T7 riser 27 4263067.29 636485.37 1731.75 175.49 223.71 T7 tread 28 4263079.11 636511.66 1732.48 176.23 252.01 T7 tread 29 4263094.78 636557.22 1733.69 177.44 298.59 T7 tread 30 4263107.92 636588.12 1734.08 177.83 331.68 T7 tread 31 4263120.73 636619.17 1732.03 175.78 364.74 T7 tread 32 4263122.39 636624.60 1730.54 174.28 370.21 T7 riser 33 4263124.35 636632.73 1726.93 170.68 378.13 T7 riser 34 4263049.04 636652.57 1732.00 175.75 358.02 T7 strath 35 4263039.26 636658.91 1729.70 173.45 359.19 T7 strath 36 4262989.72 636682.45 1730.67 174.42 359.56 T7 strath 37 4262961.14 636702.46 1729.36 173.11 366.59 T7 strath 38 4263121.13 636657.39 1716.74 160.49 397.38 bedrock 39 4263142.70 636666.35 1714.07 157.81 416.35 colluvium 40 4263146.57 636671.92 1712.29 156.04 423.11 colluvium 41 4263146.02 636695.28 1707.90 151.65 442.59 colluvium 42 4263151.47 636703.65 1706.34 150.09 452.55 colluvium 43 4263154.44 636708.87 1704.83 148.57 458.54 bedrock 44 4263176.15 636723.40 1701.33 145.08 482.38 bedrock 45 4263179.04 636724.92 1699.31 143.06 485.22 bedrock 46 4263183.20 636731.06 1697.42 141.16 492.62 bedrock 47 4263188.39 636744.48 1700.08 143.83 506.71 bedrock 48 4263192.99 636748.30 1701.75 145.49 512.41 T6 strath 49 4263198.72 636752.64 1702.98 146.73 519.14 T6 riser 50 4263203.39 636755.68 1703.49 147.24 524.22 T6 tread 51 4263209.51 636758.84 1702.58 146.32 530.19 T6 riser 52 4263212.70 636762.47 1701.44 145.19 534.97 T6 strath 53 4263218.96 636778.93 1699.04 142.78 552.19 bedrock 54 4263232.24 636803.77 1697.85 141.60 580.24 bedrock 55 4263238.94 636812.78 1695.33 139.08 591.44 bedrock 56 4263247.84 636833.51 1695.16 138.91 613.65 bedrock 57 4263257.23 636841.86 1693.58 137.33 625.77 bedrock 58 4263304.03 636974.37 1602.77 46.52 761.21 eolian sand 59 4263326.05 637008.91 1597.54 41.29 801.81 eolian sand 60 4263334.78 637022.00 1596.45 40.20 817.42 bedrock 61 4263349.89 637032.59 1586.07 29.82 834.85 bedrock 62 4263778.00 636854.13 1601.96 45.71 796.06 T4y strath

219 63 4263788.35 636846.91 1601.31 45.06 791.34 T4y strath 64 4263777.80 636844.10 1602.01 45.76 789.49 T4y tread 65 4263364.22 637050.26 1562.82 6.56 857.48 eolian sand 66 4263368.60 637054.06 1558.10 1.85 863.13 sandy alluvium 67 4263379.62 637044.78 1534.89 -21.36 862.15 sandy alluvium 68 4263378.28 637055.76 1558.55 2.30 870.24 sandy alluvium 69 4263378.29 637061.96 1558.13 1.88 875.21 T3-T1 tread 70 4263384.97 637075.11 1557.39 1.14 889.67 T3-T1 tread 71 4263400.44 637083.60 1557.47 1.22 905.70 T3-T1 tread 72 4263407.45 637091.93 1556.90 0.65 916.54 T3-T1 tread 73 4263409.37 637093.17 1556.20 -0.05 918.68 T3-T1 tread 74 4263419.49 637108.90 1555.35 -0.90 937.29 T3-T1 tread 75 4263428.96 637114.05 1555.52 -0.73 947.06 T3-T1 tread 76 4263436.67 637121.01 1555.45 -0.80 957.21 T3-T1 tread 77 4263433.96 637122.18 1556.25 0.00 956.55 water surface 78 4263440.42 637124.90 1556.03 -0.22 962.56 sandy alluvium 79 4263441.51 637127.21 1559.85 3.60 965.06 sandy alluvium 80 4263445.30 637133.62 1554.45 -1.81 972.47 sandy alluvium 81 4263444.61 637130.89 1561.73 5.47 969.86 bedrock 82 4263451.07 637151.33 1602.09 45.84 990.07 bedrock 83 4263456.75 637153.73 1604.64 48.39 995.41 bedrock 84 4263458.66 637158.07 1606.65 50.39 1000.02 T4y strath 85 4263426.84 637206.86 1608.62 52.37 1012.89 T4y strath 86 4263452.43 637186.77 1609.92 53.67 1017.90 IRSL site 87 4263461.27 637160.14 1607.97 51.72 1003.24 T4y riser 88 4263465.98 637163.35 1609.86 53.61 1008.62 T4y tread 89 4263469.52 637168.37 1610.45 54.20 1014.76 T4y tread 90 4263476.67 637175.71 1609.72 53.46 1024.91 T4y tread 91 4263478.55 637179.26 1608.50 52.25 1028.87 T4y riser 92 4263481.11 637182.89 1607.08 50.82 1033.31 T4y strath 93 4263488.67 637191.77 1603.18 46.93 1044.94 colluvium 94 4263496.10 637206.16 1598.53 42.28 1060.88 bedrock 95 4263506.40 637219.13 1594.38 38.12 1077.43 bedrock 96 4263519.24 637227.97 1591.36 35.11 1092.18 bedrock 97 4263535.12 637247.78 1600.04 43.79 1117.52 bedrock 98 4263429.56 637227.05 1611.11 54.85 1026.74 T4y strath 99 4263422.42 637269.81 1615.21 58.95 1040.10 T4o strath 100 4263394.90 637278.75 1614.93 58.68 1019.20 T4o strath 101 4263370.25 637296.71 1616.74 60.49 1002.64 T4o strath 102 4263358.84 637306.38 1617.86 61.60 994.94 T4o strath 103 4263359.88 637308.23 1618.79 62.53 996.13 T4o riser 104 4263362.46 637312.50 1619.70 63.44 998.96 T4o tread 105 4263368.00 637322.87 1620.00 63.75 1004.95 T4o tread 106 4263381.91 637348.32 1619.53 63.28 1017.95 T4o tread 107 4263379.00 637352.20 1619.05 62.79 1015.57 T4o riser 108 4263379.02 637355.11 1618.23 61.98 1015.56 T4o strath 109 4263388.03 637369.37 1613.00 56.75 1022.10 colluvium 110 4263399.39 637392.69 1608.30 52.04 1027.92 bedrock

220

Figure B.7. Pack1 GPS transect (RK 20.7). Cross section, view downstream.

TABLE B.6. SURVEY DATA FOR PACK1 GPS TRANSECT (RK 20.7) Survey type: RTK Raw Adjusted Height Distance elevation elevation above along Northing Easting ASL ASL grade transect Point (m) (m) (m) (m)* (m) (m) Notes base 4258387.53 637034.60 1579.84 1605.84 47.48 N.A. base 1 4258337.43 636186.44 1529.70 1555.70 -2.66 0.00 T1i tread 2 4258363.45 636202.83 1529.71 1555.71 -2.65 30.74 T1i tread 3 4258375.96 636210.41 1529.02 1555.02 -3.34 45.34 T1y lobe 4 4258384.50 636215.48 1530.96 1556.96 -1.39 55.26 T1y lobe 5 4258400.00 636224.97 1531.33 1557.33 -1.03 73.38 T1y lobe 6 4258406.25 636228.26 1530.32 1556.32 -2.03 80.41 T1y lobe 7 4258412.89 636231.02 1531.41 1557.41 -0.95 87.48 T1y lobe 8 4258443.87 636254.59 1532.66 1558.66 0.30 126.15 T1y lobe 9 4258452.24 636259.57 1532.33 1558.33 -0.03 135.84 T1y lobe 10 4258484.76 636278.97 1533.06 1559.06 0.71 173.46 T1y lobe 11 4258497.84 636288.54 1533.77 1559.77 1.41 189.56 T1y lobe 12 4258510.68 636298.62 1533.67 1559.67 1.31 205.76 T1y lobe 13 4258511.28 636299.94 1532.39 1558.39 0.03 207.01 modern alluvium 14 4258515.81 636306.36 1532.36 1558.36 0.00 214.39 modern alluvium 15 4258516.68 636307.56 1532.99 1558.99 0.63 215.79 T1y lobe 16 4258520.74 636312.13 1532.79 1558.79 0.43 221.74 T1y lobe 17 4258527.93 636319.64 1533.82 1559.82 1.46 231.95 T1y lobe 18 4258535.05 636326.80 1532.98 1558.98 0.62 241.91 T1y lobe 19 4258538.62 636330.66 1533.88 1559.88 1.52 247.07 T1y lobe 20 4258553.79 636344.33 1533.60 1559.60 1.25 267.36 T1y lobe 21 4258563.43 636354.42 1533.77 1559.77 1.41 281.14 T1y lobe 22 4258570.41 636361.66 1534.52 1560.52 2.16 291.08 T1y lobe 23 4258579.48 636369.05 1533.60 1559.60 1.24 302.75 T1y lobe 24 4258591.43 636380.04 1534.71 1560.71 2.35 318.91 T1y lobe 25 4258600.43 636388.70 1533.84 1559.84 1.48 331.33 T1y lobe 26 4258607.33 636392.47 1534.90 1560.90 2.54 339.03 T1y lobe 27 4258624.89 636410.80 1534.84 1560.84 2.48 364.19 T1y lobe 28 4258629.95 636415.91 1534.12 1560.12 1.76 371.34 T1y lobe 29 4258637.96 636425.52 1534.57 1560.57 2.22 383.63 T1y lobe

221 30 4258650.47 636438.50 1534.45 1560.45 2.09 401.55 T1y lobe 31 4258655.14 636445.33 1534.86 1560.86 2.50 409.54 slopewash 32 4258656.90 636447.91 1535.63 1561.63 3.27 412.55 T3 riser 33 4258659.42 636452.12 1537.88 1563.88 5.53 417.22 T3 riser 34 4258661.69 636457.28 1540.76 1566.76 8.40 422.32 T3 riser 35 4258663.31 636461.36 1543.06 1569.06 10.70 426.22 T3 riser 36 4258664.35 636462.99 1543.66 1569.66 11.30 428.08 T3 tread 37 4258670.95 636472.14 1544.67 1570.67 12.31 439.08 T3 tread 38 4258688.87 636489.88 1545.87 1571.87 13.51 464.26 T3 tread 39 4258705.71 636506.83 1546.49 1572.49 14.13 488.09 T3 tread 40 4258716.92 636516.24 1545.99 1571.99 13.64 502.73 T3 tread 41 4258722.41 636520.68 1545.34 1571.34 12.98 509.78 T3 tread 42 4258727.78 636525.78 1544.38 1570.38 12.02 517.18 T3 tread 43 4258736.56 636533.80 1541.92 1567.92 9.56 529.06 T3 riser 44 4258742.18 636538.77 1540.34 1566.34 7.98 536.56 colluvium 45 4258748.61 636543.38 1538.88 1564.88 6.52 544.42 colluvium 46 4258754.73 636546.89 1537.92 1563.92 5.56 551.32 colluvium 47 4258756.05 636547.77 1538.25 1564.25 5.89 552.89 T3 riser 48 4258760.73 636551.42 1540.12 1566.12 7.76 558.81 T3 riser 49 4258767.72 636556.85 1542.95 1568.95 10.59 567.63 T3 riser 50 4258774.96 636563.23 1545.93 1571.93 13.57 577.28 T3 riser 51 4258784.95 636571.48 1549.54 1575.54 17.18 590.21 T3 riser 52 4258795.29 636578.95 1552.85 1578.85 20.50 602.88 T3 riser 53 4258804.00 636586.08 1555.13 1581.13 22.77 614.11 T3 riser 54 4258807.71 636589.34 1555.98 1581.98 23.62 619.04 T3 tread 55 4258812.73 636594.13 1556.89 1582.89 24.53 625.98 T3 tread 56 4258820.46 636601.37 1557.87 1583.87 25.51 636.56 T3 tread 57 4258828.60 636610.33 1558.45 1584.45 26.09 648.61 T3 tread 58 4258839.69 636620.74 1558.93 1584.93 26.57 663.82 T3 tread 59 4258851.46 636632.68 1560.19 1586.19 27.83 680.55 T3 tread 60 4258861.75 636643.48 1561.27 1587.27 28.92 695.43 T3 tread 61 4258876.19 636654.78 1561.80 1587.80 29.44 713.71 T3 tread 62 4258885.54 636662.61 1561.14 1587.14 28.78 725.88 T3 tread 63 4258893.70 636670.33 1559.67 1585.67 27.31 737.10 T3 riser 64 4258905.80 636680.27 1556.52 1582.52 24.16 752.72 T3 riser 65 4258918.33 636693.76 1553.69 1579.69 21.33 771.08 T3 riser 66 4258928.00 636702.11 1551.04 1577.04 18.69 783.83 T3 riser 67 4258935.61 636711.37 1549.04 1575.04 16.68 795.72 T3 riser 68 4258947.07 636723.72 1544.05 1570.05 11.69 812.53 T3 riser 69 4258951.11 636727.29 1542.18 1568.18 9.82 817.91 colluvium 70 4258956.60 636732.13 1540.54 1566.54 8.18 825.22 colluvium 71 4258969.18 636743.18 1538.25 1564.25 5.89 841.96 T1 tread 72 4258975.32 636748.58 1536.77 1562.77 4.41 850.13 T1 tread 73 4258988.74 636759.36 1537.84 1563.84 5.48 867.26 T1 tread 74 4259012.20 636783.50 1537.84 1563.84 5.48 900.90 T1 tread 75 4259018.88 636790.37 1537.23 1563.23 4.87 910.48 T1 tread 76 4259040.37 636812.53 1539.39 1565.39 7.03 941.32 T1 tread 77 4259079.14 636848.87 1539.05 1565.05 6.70 994.45 T1 tread *26m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see App. B intro).

222

Figure B.8. Pack2 GPS transect (RK 21.9). Cross section, view downstream.

TABLE B.7. SURVEY DATA FOR PACK2 GPS TRANSECT (RK 21.9) Survey type: RTK Raw Corrected Adjusted Ht. Dist. elevation elevation elevation above along Northing Easting ASL ASL ASL grade transect Point (m) (m) (m) (m)* (m)† (m) (m) Notes Day 1 base1 4258387.53 637034.60 1579.84 1579.84 1601.84 7.42 N.A. base 1 4258187.95 637548.04 1581.13 1581.13 1603.13 8.71 365.02 T1 tread 2 4258160.26 637517.91 1580.83 1580.83 1602.83 8.41 324.10 T1 tread 3 4258153.40 637509.55 1579.89 1579.89 1601.89 7.47 313.30 T1 tread 4 4258144.84 637499.91 1580.82 1580.82 1602.82 8.40 300.42 colluvium 5 4258132.64 637489.59 1582.75 1582.75 1604.75 10.33 284.55 colluvium 6 4258127.23 637484.09 1585.27 1585.27 1607.27 12.85 276.84 colluvium 7 4258120.84 637477.87 1589.08 1589.08 1611.08 16.66 267.93 colluvium 8 4258116.47 637473.13 1591.98 1591.98 1613.98 19.56 261.48 bedrock 9 4258114.38 637470.54 1594.16 1594.16 1616.16 21.74 258.16 bedrock 10 4258111.46 637467.32 1595.71 1595.71 1617.71 23.29 253.82 T3 riser 11 4258106.60 637462.28 1598.27 1598.27 1620.27 25.85 246.82 T3 riser 12 4258101.14 637457.45 1600.26 1600.26 1622.26 27.84 239.55 T3 riser 13 4258095.37 637450.18 1601.71 1601.71 1623.71 29.29 230.31 T3 tread 14 4258083.33 637438.46 1602.87 1602.87 1624.87 30.45 213.52 T3 tread 15 4258071.19 637428.53 1603.23 1603.23 1625.23 30.81 197.94 T3 tread 16 4258059.24 637420.42 1603.11 1603.11 1625.11 30.69 183.79 T3 tread 17 4258051.69 637413.30 1602.91 1602.91 1624.91 30.49 173.42 T3 tread 18 4258050.61 637412.71 1602.62 1602.62 1624.62 30.20 172.24 T3 riser 19 4258039.61 637404.53 1594.43 1594.43 1616.43 22.01 158.70 T3 riser 20 4258027.77 637402.56 1586.49 1586.49 1608.49 14.07 149.02 colluvium 21 4258018.76 637373.13 1579.31 1579.31 1601.31 6.89 121.67 colluvium 22 4258016.30 637343.18 1576.77 1576.77 1598.77 4.35 98.88 colluvium 23 4257999.57 637337.97 1575.56 1575.56 1597.56 3.14 83.28 T1y tread

Day 2 base2 4257972.80 637564.29 1588.14 1585.89 1607.89 13.47 N.A. base 24 4257999.37 637336.78 1577.81 1575.56 1597.56 3.14 82.31 T1y tread

223 25 4257993.53 637336.13 1577.64 1575.39 1597.39 2.97 77.68 T1y tread 26 4257990.07 637335.67 1577.11 1574.86 1596.86 2.44 74.89 T1y riser 27 4257987.83 637335.85 1576.20 1573.95 1595.95 1.53 73.43 T1yy tread 28 4257978.33 637334.98 1575.96 1573.71 1595.71 1.29 66.14 T1yy tread 29 4257970.04 637334.03 1575.79 1573.54 1595.54 1.12 59.76 T1yy tread 30 4257968.99 637334.72 1574.96 1572.71 1594.71 0.29 59.56 modern alluvium 31 4257964.43 637335.28 1574.67 1572.42 1594.42 0.00 56.94 modern alluvium 32 4257962.54 637336.19 1577.51 1575.26 1597.26 2.84 56.38 T1y riser 33 4257960.11 637335.91 1578.43 1576.18 1598.18 3.76 54.58 T1y tread 34 4257950.55 637334.18 1577.93 1575.68 1597.68 3.26 47.23 T1y tread 35 4257945.90 637333.50 1578.38 1576.13 1598.13 3.71 43.88 T1y tread 36 4257940.14 637332.83 1577.77 1575.52 1597.52 3.10 39.98 T1y riser 37 4257936.97 637331.86 1576.84 1574.59 1596.59 2.17 37.41 T1y riser 38 4257934.19 637331.30 1576.71 1574.46 1596.46 2.04 35.42 T1y riser 39 4257932.46 637330.48 1576.21 1573.96 1595.96 1.54 33.81 modern alluvium 40 4257928.44 637328.71 1575.29 1573.04 1595.04 0.63 30.24 modern alluvium 41 4257925.42 637327.94 1576.59 1574.34 1596.34 1.92 28.07 T1i riser 42 4257921.41 637327.14 1579.54 1577.29 1599.29 4.87 25.43 T1i riser 43 4257918.56 637326.30 1580.31 1578.06 1600.06 5.64 23.38 T1i tread 44 4257906.42 637324.37 1580.74 1578.49 1600.49 6.07 16.64 T1i tread 45 4257853.90 637314.07 1580.46 1578.21 1600.21 5.79 0.00 T1i tread *Drift correction applied to elevation of points 24-45 to account for differences in a repeat survey point (23 and 24); points 1-23 (with base1) were surveyed one day, points 24-45 (with base2) were surveyed the next day due to RTK battery limitations. †22m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see Appendix B introduction).

224

Figure B.9. Pack3 GPS transect (RK 23.5). Cross section, view downstream.

TABLE B.8. SURVEY DATA FOR PACK3 GPS TRANSECT (RK 23.5) Survey type: RTK Raw Adjusted Height Distance elevation elevation above along Northing Easting ASL ASL grade transect Point (m) (m) (m) (m)* (m) (m) Notes base 4256890.27 638207.78 1670.69 1692.69 50.49 N.A. base 1 4257604.61 638585.40 1635.88 1657.88 15.68 1132.68 bedrock 2 4257588.94 638576.92 1632.31 1654.31 12.11 1114.86 colluvium 3 4257567.95 638568.24 1630.40 1652.40 10.20 1092.20 T1o tread 4 4257553.74 638561.45 1629.53 1651.53 9.33 1076.46 T1o tread 5 4257526.21 638549.14 1627.88 1649.88 7.68 1046.31 T1o tread 6 4257523.91 638547.74 1627.40 1649.40 7.20 1043.63 T1o riser 7 4257520.64 638545.35 1626.71 1648.71 6.50 1039.66 T1y tread 8 4257491.33 638530.40 1625.76 1647.76 5.55 1006.77 T1y tread 9 4257482.21 638525.83 1624.64 1646.64 4.43 996.57 T1y tread 10 4257479.73 638524.71 1622.88 1644.88 2.68 993.85 T1y riser 11 4257477.15 638523.01 1621.42 1643.42 1.21 990.78 modern alluvium 12 4257470.19 638519.31 1620.99 1642.99 0.79 982.91 modern alluvium 13 4257466.21 638518.44 1620.21 1642.21 0.00 978.95 water surface 14 4257464.10 638518.04 1620.27 1642.27 0.07 976.88 water surface 15 4257462.55 638517.90 1620.54 1642.54 0.34 975.43 modern alluvium 16 4257461.29 638517.60 1621.94 1643.94 1.74 974.17 T1yy tread 17 4257451.89 638514.55 1622.38 1644.38 2.17 964.38 T1yy tread 18 4257431.75 638506.33 1623.22 1645.22 3.01 942.68 T1yy tread 19 4257428.60 638504.69 1625.02 1647.02 4.81 939.13 T1o riser 20 4257425.65 638502.86 1627.12 1649.12 6.91 935.68 T1o riser 21 4257419.63 638499.77 1628.08 1650.08 7.87 928.93 T1o tread 22 4257396.76 638488.19 1628.28 1650.28 8.08 903.31 T1o tread 23 4257381.52 638480.76 1628.10 1650.10 7.90 886.38 T1o tread 24 4257351.14 638466.50 1626.33 1648.33 6.13 852.86 T1y tread 25 4257338.77 638459.20 1626.92 1648.92 6.72 838.54 T1y tread

225 26 4257320.13 638447.22 1626.16 1648.16 5.96 816.53 T1y tread 27 4257285.41 638427.39 1625.81 1647.81 5.60 776.62 T1y tread 28 4257274.14 638420.94 1624.77 1646.77 4.57 763.65 T1y tread 29 4257271.35 638419.25 1624.98 1646.98 4.78 760.40 T1y tread 30 4257266.93 638416.94 1624.43 1646.43 4.22 755.41 T1y tread 31 4257253.89 638410.60 1625.14 1647.14 4.94 740.92 T1y tread 32 4257248.75 638408.34 1625.85 1647.85 5.64 735.32 T1y tread 33 4257239.58 638403.23 1625.91 1647.91 5.70 724.84 T1y tread 34 4257232.40 638400.50 1626.71 1648.71 6.50 717.21 T1y tread 35 4257223.07 638395.27 1626.17 1648.17 5.97 706.54 T1y tread 36 4257215.29 638390.85 1627.23 1649.23 7.03 697.60 T1y tread 37 4257211.06 638388.68 1628.35 1650.35 8.14 692.85 T2 riser 38 4257205.09 638386.51 1631.53 1653.53 11.32 686.55 T2 riser 39 4257199.78 638383.84 1634.61 1656.61 14.41 680.62 T2 riser 40 4257197.33 638382.21 1635.86 1657.86 15.65 677.69 T2 riser 41 4257195.32 638381.00 1636.36 1658.36 16.15 675.35 T2 tread 42 4257193.11 638379.41 1636.42 1658.42 16.21 672.66 T2 tread 43 4257188.47 638376.65 1636.09 1658.09 15.88 667.26 T2 tread 44 4257183.46 638374.26 1635.47 1657.47 15.26 661.72 T2 riser 45 4257176.91 638370.16 1633.94 1655.94 13.74 654.03 T2 riser 46 4257165.29 638364.94 1631.04 1653.04 10.84 641.32 T2 riser 47 4257157.69 638360.75 1629.00 1651.00 8.80 632.65 T2 riser 48 4257149.97 638356.85 1626.77 1648.77 6.57 624.01 colluvium 49 4257141.24 638352.38 1624.92 1646.92 4.72 614.21 alluvial sand 50 4257126.06 638344.90 1624.24 1646.24 4.03 597.31 alluvial sand 51 4257117.54 638340.62 1624.49 1646.49 4.29 587.78 alluvial sand 52 4257112.93 638337.82 1624.07 1646.07 3.87 582.40 alluvial sand 53 4257110.97 638337.71 1625.81 1647.81 5.61 580.63 alluvial sand 54 4257094.44 638330.56 1627.59 1649.59 7.39 562.70 alluvial sand 55 4257082.63 638323.99 1629.56 1651.56 9.36 549.20 T3 riser 56 4257066.51 638316.54 1633.43 1655.43 13.23 531.51 T3 riser 57 4257050.20 638308.19 1637.65 1659.65 17.44 513.22 T3 riser 58 4257038.13 638302.52 1641.19 1663.19 20.99 499.92 T3 riser 59 4257027.01 638298.52 1645.13 1667.13 24.92 488.27 T3 riser 60 4257010.97 638291.25 1650.62 1672.62 30.42 470.75 T3 riser 61 4256997.21 638284.78 1654.52 1676.52 34.31 455.61 T3 riser 62 4256985.37 638276.69 1656.68 1678.68 36.48 441.36 T3 riser 63 4256971.59 638266.76 1658.39 1680.39 38.18 424.50 T3 riser 64 4256953.84 638252.17 1662.39 1684.39 42.18 401.88 T3 riser 65 4256943.96 638242.74 1664.74 1686.74 44.54 388.67 T3 riser 66 4256934.60 638235.01 1666.46 1688.46 46.25 376.72 T3 riser 67 4256919.11 638221.98 1668.47 1690.47 48.26 356.86 T3 riser 68 4256910.85 638211.58 1669.70 1691.70 49.50 344.61 T3 tread 69 4256902.12 638203.18 1670.24 1692.24 50.04 332.94 T3 tread 70 4256884.64 638194.53 1669.77 1691.77 49.56 313.45 T3 tread 71 4256867.10 638185.66 1668.49 1690.49 48.29 293.81 T3 tread 72 4256846.82 638172.99 1666.64 1688.64 46.43 269.94 T3 tread 73 4256826.88 638162.03 1665.46 1687.46 45.25 247.19 T3 tread 74 4256808.69 638151.19 1664.84 1686.84 44.63 226.05 T3 tread

226 75 4256726.90 638172.93 1666.44 1688.44 46.23 171.71 T3 tread 76 4256713.80 638164.76 1665.39 1687.39 45.18 156.91 T3 tread 77 4256706.28 638158.31 1663.85 1685.85 43.65 147.23 T3 riser 78 4256697.97 638150.45 1660.78 1682.78 40.57 136.01 T3 riser 79 4256689.00 638141.97 1656.50 1678.50 36.29 123.89 T3 riser 80 4256680.86 638135.13 1652.26 1674.26 32.05 113.53 T3 riser 81 4256669.41 638128.53 1646.25 1668.25 26.04 101.30 T3 riser 82 4256661.79 638127.49 1643.18 1665.18 22.97 95.93 T3 riser 83 4256654.62 638125.35 1639.56 1661.56 19.36 90.24 T3 riser 84 4256647.59 638123.32 1636.28 1658.28 16.07 84.99 T3 strath 85 4256643.61 638120.60 1632.74 1654.74 12.54 80.81 bedrock 86 4256639.75 638119.24 1629.17 1651.17 8.97 77.93 bedrock 87 4256635.65 638115.05 1626.08 1648.08 5.88 72.56 colluvium 88 4256625.72 638103.57 1622.45 1644.45 2.25 58.66 colluvium 89 4256622.88 638090.17 1618.79 1640.79 -1.41 45.45 colluvium 90 4256631.77 638067.13 1615.45 1637.45 -4.75 31.17 colluvium 91 4256629.16 638061.57 1614.60 1636.60 -5.60 25.68 P1 tread 92 4256620.73 638052.85 1613.85 1635.85 -6.36 13.92 P1 tread 93 4256614.46 638047.70 1611.91 1633.91 -8.29 6.26 bedrock 94 4256608.30 638046.59 1614.38 1636.38 -5.83 0.00 bedrock *22m added to all raw RTK elevations to make comparable with DEM and handheld GPS (see Appendix B introduction).

Figure B.10. Pack4 GPS transect (RK 25.4). Cross section, view downstream. 227

228 TABLE B.9. SURVEY DATA FOR PACK4 GPS TRANSECT (RK 25.4) Survey type: handheld Height Distance Elevation above along Northing Easting ASL grade transect Point (m) (m) (m) (m) (m) Notes 1 4257101.36 640058.32 1711.88 13.17 1384.28 eolian sand 2 4257051.97 640048.43 1710.58 11.87 1333.91 T1o tread 3 4256991.97 640007.39 1709.10 10.39 1263.94 T1o tread 4 4256957.52 640012.75 1709.26 10.55 1233.51 T1o tread 5 4256935.97 640012.37 1708.85 10.14 1212.69 T1o tread 6 4256926.53 640012.66 1706.86 8.15 1203.64 T1o riser 7 4256916.69 640010.43 1702.99 4.28 1193.55 sandy alluvium 8 4256905.89 640007.31 1701.79 3.08 1182.33 sandy alluvium 9 4256904.44 640007.48 1699.92 1.21 1180.96 modern channel 10 4256897.69 640007.98 1698.71 0.00 1174.52 water surface 11 4256895.59 640010.36 1702.98 4.27 1173.02 water surface 12 4256892.31 640009.65 1702.91 4.19 1169.67 modern alluvium 13 4256881.96 640003.28 1701.38 2.67 1158.14 modern alluvium 14 4256879.35 640003.07 1701.37 2.66 1155.55 T1y riser 15 4256875.98 640003.24 1702.87 4.16 1152.32 T1y tread 16 4256861.95 640000.17 1702.45 3.74 1137.96 T1y tread 17 4256849.45 639975.93 1701.75 3.04 1119.70 T1y tread 18 4256847.97 639971.40 1704.51 5.80 1116.99 T1o riser 19 4256845.46 639969.42 1706.58 7.87 1114.02 T1o tread 20 4256821.20 639966.90 1706.55 7.84 1090.17 T1o tread 21 4256801.55 639968.67 1706.84 8.13 1071.78 T1o tread 22 4256782.74 639964.04 1706.00 7.28 1052.47 T1o tread 23 4256768.82 639962.62 1704.71 6.00 1038.68 modern alluvium 24 4256763.36 639961.27 1705.12 6.41 1033.07 modern alluvium 25 4256760.66 639961.33 1707.51 8.80 1030.48 T1o tread 26 4256735.47 639965.63 1708.18 9.46 1007.05 T1o tread 27 4256722.95 639968.07 1709.05 10.34 995.39 P1 tread 28 4256684.56 639963.71 1711.28 12.57 956.97 P1 tread 29 4256629.84 639976.00 1714.92 16.21 905.66 P1 tread 30 4256589.79 639984.41 1718.52 19.81 867.72 P1 tread 31 4256585.14 639984.11 1717.06 18.35 863.09 modern alluvium 32 4256576.56 639983.08 1718.11 19.40 854.47 modern alluvium 33 4256571.30 639982.24 1718.97 20.25 849.15 P1 tread 34 4256566.78 639981.34 1719.13 20.42 844.55 P1 tread 35 4256560.48 639980.45 1721.79 23.08 838.20 T2 riser 36 4256550.97 639980.99 1726.31 27.60 828.93 T2 riser 37 4256544.50 639979.51 1728.95 30.24 822.32 T2 riser 38 4256541.71 639978.86 1729.39 30.68 819.47 T2 riser 39 4256528.17 639976.63 1730.39 31.68 805.76 T2 riser 40 4256513.62 639972.28 1732.60 33.89 790.70 T2 tread 41 4256478.97 639963.05 1734.51 35.80 755.00 T2 tread 42 4256451.75 639961.69 1735.22 36.50 728.03 T2 tread 43 4256399.84 639961.58 1739.65 40.94 677.10 P2 tread

229 44 4256323.84 639955.07 1745.51 46.80 601.54 P2 tread 45 4256235.64 639942.71 1751.83 53.12 513.12 P2 tread 46 4256211.87 639937.24 1753.74 55.03 488.85 P2 tread 47 4256197.70 639934.27 1754.21 55.50 474.45 P2 tread 48 4256176.21 639930.28 1751.56 52.85 452.73 modern alluvium 49 4256128.74 639915.54 1758.48 59.77 403.36 P2 tread 50 4256096.07 639910.12 1761.08 62.37 370.51 P2 tread 51 4256086.39 639909.47 1760.99 62.28 361.02 P2 riser 52 4256078.23 639908.43 1759.30 60.59 352.91 P2 riser 53 4256070.98 639906.53 1757.19 58.48 345.45 modern alluvium 54 4256058.68 639902.02 1760.89 62.18 332.47 P2y tread 55 4256042.92 639899.37 1761.42 62.71 316.64 P2y tread 56 4256025.20 639895.91 1763.97 65.26 298.73 P2y tread 57 4256013.26 639893.36 1762.97 64.26 286.62 bedrock 58 4256009.65 639892.44 1762.36 63.65 282.91 modern alluvium 59 4255999.42 639886.28 1765.80 67.09 271.43 P3 riser 60 4255986.62 639878.85 1771.25 72.54 257.16 P3 riser 61 4255949.41 639870.19 1779.82 81.11 219.13 P3 tread 62 4255925.28 639861.47 1783.12 84.41 193.61 P3 tread 63 4255909.22 639855.58 1784.48 85.77 176.58 P3 tread 64 4255900.72 639851.75 1787.94 89.23 167.37 P3 tread 65 4255887.53 639841.69 1791.22 92.51 151.99 P3 tread 66 4255861.23 639829.35 1796.85 98.14 123.47 P3 tread 67 4255842.77 639821.07 1800.68 101.97 103.62 P3 tread 68 4255839.28 639819.60 1801.19 102.48 99.89 P3 tread 69 4255823.16 639812.15 1797.63 98.92 82.59 P3 riser 70 4255793.60 639797.92 1781.19 82.48 51.15 P3 riser 71 4255778.54 639789.54 1774.28 75.57 35.44 colluvium 72 4255763.73 639781.38 1770.65 71.94 20.66 P2 tread 73 4255736.86 639758.17 1765.94 67.23 1.01 P2 tread 74 4255734.41 639756.13 1764.58 65.87 0.00 P2 riser 75 4255733.40 639753.15 1761.35 62.64 0.83 P2 riser 76 4255732.17 639751.52 1759.99 61.28 0.82 P2 strath

230

Appendix C: Supplemental terrain analysis data

231 Overview

Supplemental catchment metrics and statistics reported include: basin area; minimum, maximum, and mean elevation; total relief; hypsometric integral (HI); ratio of volume to area (RVA); and mean, median, and standard deviation of slope. Statistics were calculated for Courthouse Wash, Grandstaff Canyon, Mill Creek (combined North and South Mill), and Pack Creek (combined Brumley and upper Pack). The Mill and Pack creek catchments were clipped at the Mill-Pack confluence rather than the Colorado River. Analyses were performed using mosaicked 10 m digital elevation models from the USGS National Elevation Dataset in ArcGIS 10.3.1 and Microsoft Excel 2016. Elevation and slope statistics were calculated using built-in ArcGIS functions. HI was calculated by dividing the difference between the mean and minimum catchment elevations by the catchment’s total relief. RVA was calculated according to the procedure of Frankel and Pazzaglia (2005). Slope histograms were generated by calculating the slope between neighboring cells for each pixel in the catchment using the Slope tool in the ArcGIS 3D Analyst toolbox. These values were exported to Microsoft Excel and placed in 2° bins to create histograms. Finally, the Python script used for batch-processing in the generation of longitudinal profiles and slope-area plots presented in Chapter 3 is included.

TABLE C.1. SUPPLEMENTAL CATCHMENT STATISTICS Elevation Slope Area (m) (°) Catchment (km2) RVA Min Max Mean Relief HI Med Mean Std dev Courthouse Wash 421 59 1207 1847 1465 640 0.40 4.20 7.68 9.82 Grandstaff Canyon 52 56 1208 2331 1623 1123 0.37 9.99 14.14 12.46 Mill Creek 199 56 1220 3856 2164 2637 0.36 11.05 15.22 12.52 Pack Creek 167 153 1220 3810 2034 2590 0.31 11.39 14.63 12.44

232 A Courthouse Wash 1200000

1000000

800000

600000 # pixels 400000

200000

0 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 slope bins, degrees (upper bin value)

B Grandstaff Canyon 70000

60000

50000

40000

# pixels 30000

20000

10000

0 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 slope bins, degrees (upper bin value)

C Mill Creek 250000

200000

150000

# pixels 100000

50000

0 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 slope bins, degrees (upper bin value) Figure C.1. continues

233 D Pack Creek 250000

200000

150000

# pixels 100000

50000

0 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 slope bins, degrees (upper bin value)

Figure C.1. Slope distribution histograms. Catchments are Courthouse Wash (A), Grandstaff Canyon (B), Mill Creek (C), and Pack Creek (D). Note that Mill and Pack creeks, which have parts of their catchments in the Moab-Spanish Valley salt graben, are more positively-skewed than Courthouse Wash and Grandstaff Canyon, which have catchments almost entirely outside the graben.

Python batch-processing script for vertically-sampled longitudinal profiles and slope-area plots.

1. # Generates a CSV of x, z, and a values for slope- area plots and long profiles; points extracted along elevation contours 2. 3. # Import modules 4. import arcpy 5. import os 6. import csv 7. from arcpy import env 8. 9. # Inputs 10. dem = r'F:\USU\Thesis\Terrain_analysis\Aug2017_ksn\data\slope_area_cum\clip.tif'

11. # Raw mosaicked and projected DEM 12. fac = r'F:\USU\Thesis\Terrain_analysis\Aug2017_ksn\data\slope_area_cum\fac.tif'

13. # Flow accumulation raster 14. contour = r'F:\USU\Thesis\Terrain_analysis\Aug2017_ksn\data\slope_area_cum\conto ur_12m.shp' 15. # Contour shapefile from DEM 16. 17. # Set workspace and allow files to be overwritten 18. env.workspace = r'F:\USU\Thesis\Terrain_analysis\Aug2017_ksn\data\slope_area_cum \Courthouse' 19. env.overwriteOutput = True 20. 21. # Input variables 22. strm = 'strm.shp' # Trunk stream shapefile

234 23. cellsz = 10 # x dimension of cell (m) 24. 25. # Output variables 26. strpts = 'strpts12.shp' # Points along stream generated at the speci fied vertical contour interval 27. strpts_s = 'strpts12_sin.shp' # Streampoints shapefile of type point rathe r than multipoint 28. strpts_z = 'strpts12_z.shp' # Streampoints shapefile with distance, elev ation, and area 29. strpts_z2 = 'strpts12m_z.shp' # Streampoints shapefile with distance, elev ation, and area; sorted by distance 30. strmlr = 'strmlr.shp' # Trunk stream shapefile linear referenced 31. x_table = 'x_table.dbf' # Table of distance values along the streaml ine 32. x_table_sort = 'x_table_sort.dbf' # Table of distance values along the streaml ine, sorted in ascending order 33. profile = 'profile_contour.csv' # Distance, elevation, and area CSV file 34. 35. # Define the spatial reference system for outputs 36. srs = arcpy.Describe(dem).spatialReference 37. 38. # Check out extensions 39. arcpy.CheckOutExtension('Spatial') 40. arcpy.CheckOutExtension('3D') 41. 42. # Create points where the streamline intersects contour lines and project the ou tput 43. arcpy.Intersect_analysis([strm,contour],strpts,'ALL','','POINT') 44. arcpy.DefineProjection_management(strpts,srs) 45. 46. # Convert multipoints into points 47. arcpy.MultipartToSinglepart_management(strpts,strpts_s) 48. 49. # Make streamline a route 50. arcpy.CreateRoutes_lr(strm,'GRID_CODE',strmlr,'LENGTH','','','LOWER_RIGHT') 51. # Change to 'LOWER_RIGHT' for Courthouse, 'UPPER_LEFT' for all others 52. 53. # Calculate streamwise distance for each point, and export as table 54. props = "GRID_CODE POINT DISTANCE" 55. arcpy.LocateFeaturesAlongRoutes_lr(strpts_s,strmlr,'GRID_CODE',0.1,x_table,props ) 56. 57. # Sort by distance 58. arcpy.Sort_management(x_table,x_table_sort,[['DISTANCE','ASCENDING']]) 59. 60. # Delete rows with identical contour values in table 61. arcpy.DeleteIdentical_management(x_table_sort,'CONTOUR') 62. 63. # Join streamwise distance table to points shapefile 64. arcpy.JoinField_management(strpts_s,'CONTOUR',x_table_sort,'CONTOUR','DISTANCE')

65. 66. # Delete rows with identical contour values in points shapefile 67. arcpy.DeleteIdentical_management(strpts_s,'CONTOUR') 68. 69. # Extract area from fac raster to the points shapefile 70. arcpy.sa.ExtractValuesToPoints(strpts_s, fac, strpts_z) 71. 72. # Convert number of pixels to area values 73. arcpy.AddField_management(strpts_z, 'AREA', 'DOUBLE', 15, 0)

235 74. with arcpy.da.UpdateCursor(strpts_z, ['RASTERVALU', 'AREA']) as updater: 75. for row in updater: 76. row[1] = row[0] * cellsz**2 77. updater.updateRow(row) 78. 79. # Sort by distance 80. arcpy.Sort_management(strpts_z,strpts_z2,[['DISTANCE','ASCENDING']]) 81. 82. # Set work directories with os for the CSV writer 83. work_folder = arcpy.env.workspace 84. os.chdir(work_folder) 85. 86. # Write elevation, distance, and area data from the stream points shapefile to a CSV file 87. data = [] 88. with arcpy.da.SearchCursor(strpts_z2, ['DISTANCE', 'CONTOUR', 'AREA']) as search er: 89. for row in searcher: 90. data.append(row) 91. with open(profile, 'wb') as fp: 92. writer = csv.writer(fp) 93. writer.writerow(['x (m)', 'z (m)', 'a (m^2)']) 94. writer.writerows(data) 95. 96. # Delete unneeded files 97. todel = [strpts,strpts_s,strpts_z,strmlr,x_table,x_table_sort] 98. for file in todel: 99. arcpy.Delete_management(file) 100. 101. # Check in extensions 102. arcpy.CheckInExtension('Spatial') 103. arcpy.CheckInExtension('3D') 104. 105. # Script works, but clunky, 9/25/17. Need to just get elev values from "Contour" value in strmpts_12.shp instead of extracting values from DEM once again. This would eliminate the need for strmpts_uno.shp. Well... not exactly, because still need to extract area values from fac raster. Another problem: get a cluster of points when there's slight fluctuation in elevation and cross a contour multiple times. Need to delete all the points upstream of the first instance of crossing a contour. 106. 107. # Fixed these issues 9/27/17.